Modelo AutoCura

Librerías a utilizar

library(caret) # para preprocesado (scaling), nearzerovar y findcorrelation

## Loading required package: lattice

## Loading required package: ggplot2

library(DMwR) # para el balanceo SMOTE

## Warning: package 'DMwR' was built under R version 4.0.2

## Loading required package: grid

## Registered S3 method overwritten by 'quantmod':
##   method            from
##   as.zoo.data.frame zoo

library(car) # para prueba de inflación de la varianza en la regresión

## Loading required package: carData

library(pROC) # para calcular el AUC

## Type 'citation("pROC")' for a citation.

## 
## Attaching package: 'pROC'

## The following objects are masked from 'package:stats':
## 
##     cov, smooth, var

library(ROCR) # para calcular el AUC
library(glmnet) # para regresión elastic net

## Warning: package 'glmnet' was built under R version 4.0.2

## Loading required package: Matrix

## Loaded glmnet 4.0-2

library(randomForest) # para random forest

## Warning: package 'randomForest' was built under R version 4.0.2

## randomForest 4.6-14

## Type rfNews() to see new features/changes/bug fixes.

## 
## Attaching package: 'randomForest'

## The following object is masked from 'package:ggplot2':
## 
##     margin

library(adabag) # para adaboost

## Loading required package: rpart

## Loading required package: foreach

## Loading required package: doParallel

## Loading required package: iterators

## Loading required package: parallel

Cargar y Explorar el dataset

# definir directorio de trabajo
setwd("C:/Users/licja/Downloads/Prueba BAM Javier Saravia") 
getwd()

## [1] "C:/Users/licja/Downloads/Prueba BAM Javier Saravia"

original <- read.csv("Base_entrenamiento.csv", header = TRUE)
dim(original) # cargadas 20,000 filas y 124 columnas

## [1] 20000   124

data <- original # sobre este dataset se hará el feature engineer

Omitir valores atípicos

Los valores atípicos para cada variable se identificaron visualmente a través de histogramas para luego cuantificarlos a través de un análisis de percentiles

# Para los valores atípicos detectados se colocarán acá sus respectivos subset
data <- subset(data, data$pc_transaccional <= 136903394) 
data <- subset(data, data$dmi_max_egreso_diario <= 85362328)
data <- subset(data, data$dmi_max_ingreso_diario <= 104909227.3)
data <- subset(data, data$dh_val_otros_d <= 9998335.80)
data <- subset(data, data$dmi_ingreso_total_mes <= 205163705.0)
data <- subset(data, data$dh_val_pagos_d <= 50357360.00)
data <- subset(data, data$pc_gasto_familiar <= 26582180.19)
data <- subset(data, data$pc_cuotas_pagadas <= 8156850.0)
data <- subset(data, data$dmi_egreso_total_mes <= 209829245.0)
data <- subset(data, data$dh_val_salidas <= 217011100)
data <- subset(data, data$cpc_sum_saldo <= 220905770.00)
data <- subset(data, data$cpc_saldo_sobre_ing <= 20.307294513)
data <- subset(data, data$cpc_saldo_tdc <= 43186883.96)
data <- subset(data, data$dh_cant_tras_d <= 31)
data <- subset(data, data$cpc_avg_saldo <= 106884393.50)
data <- subset(data, data$dh_val_pago_tarj_d <= 10560315.0)
data <- subset(data, data$pc_transaccional <= 92269310.0)

Exploración y Transformación de variables

• Cada variable se exploró individualmente para identificar si requiere transformaciones y/o si se puede utilizar

• Muchas variables fueron descartadas por tener solo 0’s, ser nulas, o tener una varianza cercana a 0

• Cuando la variable se considera apta para uso, se guarda en una variable independiente; posteriormente se unificarán todas esas variables

Si bien este es un ejercicio, pueden hacerse las siguientes observaciones y sugerencias:

• Muchas variables contienen solo 0’s, muchas otras tienen casi toda la frecuencia concentrada en una sola clase, lo que hace que la frecuencia sea cercana a 0

• Muchas variables son repetitivas acerca del momento del mes en qué recibió algún depósito, hizo algún pago, retiro, etc. Creo que medir la variable así no solo causará multicolinealidad con las demás variables parecidas sino además tendrá un componente estacional, volviendóse así auto-regresiva, creo que es mejor hacer una variable binaria que indique si la transacción fue en la primera o segunda quincena del mes, además estaríamos con ello incrementando la varianza.

Sugiero variables como:

• Faltan todas las socio demográficas del cliente

• Interacción del cliente con el banco: cantidad de productos, suma del saldo de productos, suma del monto de productos, antigüedad o tiempo de ser cliente, etc.

• Las transacciones en cuentas de pasivos se pueden categorizar, así tendríamos variables como: suma_educacion, suma_combustible, suma_restaurantes, suma_telefonia, etc.

• Considero que de los burós podemos sacar mucha más información: cantidad, saldo y monto de tarjetas de crédito, créditos fiduciarios y otros créditos, inclusive comportamiento en créditos comerciales y pago de servicios

• A nivel de banco y nivel de buró elaboraría muchas más variables referentes a mora: max_mora_Nmeses, cantidad_moraN_Nmeses, desv_maxMora_Nmeses, etc.

• Podemos agregar también variables sobre los canales, frecuencia y modo en qué retira su dinero: ATM, transferencias, Tarjeta de Débito, etc.

# Se crearán variables independientes ya con transformaciones
# Exceptuando las columnas que no se consideren aptas para el modelo
# Posteriormente, se creará un solo dataset con todas ellas

# Se crearán variables numéricas para usarse en los modelos lineales (reg logística y elastic net)

# Se determinará qué variables dejar evaluando que la varianza no sea cercana a 0
# Además, se eliminará las variables independientes con alta correlación entre sí
# Muchas variables tienen frecuencia alta cerca de 0 (asimetría positiva), por lo que seguramente serán descartadas

names(data)[1] # max_trim: máximo días de mora trimestre anterior

## [1] "max_trim"

hist(data$max_trim)

hist(sqrt(data$max_trim)) # variable normalizada con sqrt

max_trim <- sqrt(data$max_trim)

names(data)[2] # max_sem: máximo días de mora en el semestre anterior

## [1] "max_sem"

hist(data$max_sem)

hist(sqrt(data$max_sem))

max_sem <- data$max_sem

names(data)[3] # desv_sem: desviación estándar del máximo de moras en el semestre anterior

## [1] "desv_sem"

hist(data$desv_sem)

hist(sqrt(data$desv_sem))

desv_sem <- data$desv_sem

names(data)[4] # prom_bim: promedio del máximo de moras en el bimestre anterior

## [1] "prom_bim"

hist(data$prom_bim)

hist(sqrt(data$prom_bim))

prom_bim <- sqrt(data$prom_bim)

names(data)[5] # max_mes_anterior: días de mora máximo en el mes anterior

## [1] "max_mes_anterior"

hist(data$max_mes_anterior)

hist(sqrt(data$max_mes_anterior))

max_mes_anterior <- data$max_mes_anterior

names(data)[6] # prom_mes_anterior: Promedio de los dias de mora en el mes anterior

## [1] "prom_mes_anterior"

hist(data$prom_mes_anterior)

hist(sqrt(data$prom_mes_anterior))

prom_mes_anterior <- sqrt(data$prom_mes_anterior)

names(data)[7] # prom_sem: promedio del máximo de moras en el semestre anterior

## [1] "prom_sem"

hist(data$prom_sem)

hist(sqrt(data$prom_sem))

prom_sem <- data$prom_sem

names(data)[8] # max_bim: máximo días de mora en el bimestre anterior

## [1] "max_bim"

hist(data$max_bim)

hist(sqrt(data$max_bim))

max_bim <- data$max_bim

names(data)[9] # mejor_gestion: Mejor gestion realizada

## [1] "mejor_gestion"

table(data$mejor_gestion)

## 
##     0     5     6     7    16 
## 18031     1     3     1    10

names(data)[10] # prom_trim: promedio del máximo de moras en el trimestre anterior

## [1] "prom_trim"

hist(data$prom_trim)

hist(sqrt(data$prom_trim))

prom_trim <- sqrt(data$prom_trim)

names(data)[11] # pc_cant_moras_30_ult_12_meses: Cantidad de moras 30 en los últimos 12 meses <= 5 o vacío.

## [1] "pc_cant_moras_30_ult_12_meses"

summary(data$pc_cant_moras_30_ult_12_meses)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   0.000   0.000   0.000   1.298   1.000  12.000

table(data$pc_cant_moras_30_ult_12_meses)

## 
##     0     1     2     3     4     5     6     7     8     9    10    11    12 
## 11672  2024  1146   919   500   333   321   188   182   192   196   176   197

pc_cant_moras_30_ult_12_meses <- data$pc_cant_moras_30_ult_12_meses

names(data)[12] # desv_trim: desviación estándar del máximo de moras en el trimestre  anterior

## [1] "desv_trim"

hist(data$desv_trim)

hist(sqrt(data$desv_trim))

desv_trim <- sqrt(data$desv_trim)

names(data)[13] # nro_gestiones: Numero de gestiones realizadas

## [1] "nro_gestiones"

table(data$nro_gestiones)

## 
##     0     1     2     3     5 
## 18031     6     4     4     1

names(data)[14] # desv_bim: desviación estándar del máximo días de mora en el bimestre anterior

## [1] "desv_bim"

hist(data$desv_bim)

hist(sqrt(data$desv_bim))

desv_bim <- sqrt(data$desv_bim)

names(data)[15] # pc_cant_moras_30_ult_3_meses: Cantidad de moras 30 en los últimos 3 meses = 0 o vacío.

## [1] "pc_cant_moras_30_ult_3_meses"

summary(data$pc_cant_moras_30_ult_3_meses)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##  0.0000  0.0000  0.0000  0.4745  0.0000  3.0000

hist(data$pc_cant_moras_30_ult_3_meses)

names(data)[16] # dh_cant_entradas: cantidad de trasacciones de ingreso de dinero tuvo en el mes anterior

## [1] "dh_cant_entradas"

summary(data$dh_cant_entradas)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    0.00    0.00    5.00   11.52   17.00  139.00

hist(data$dh_cant_entradas)

hist(sqrt(data$dh_cant_entradas))

dh_cant_entradas <- sqrt(data$dh_cant_entradas)

names(data)[17] # pc_tiem_1er_prod_abierto_total: Tiempo total de producto abierto

## [1] "pc_tiem_1er_prod_abierto_total"

table(data$pc_tiem_1er_prod_abierto_total)

## 
##     0     2     3     5    11    14    16    17    18    19    21    22    24 
## 17794     6     2     5     3     1     5     1     2     1     4     1     1 
##    25    26    27    28    29    30    31    33    34    35    38    39    41 
##     2     5     2     5     4     2     5     2     4     1     2    35     2 
##    42    44    46    48    49    50    51    52    53    54    55    56    57 
##     1     3     3     2    21     4     2     2     1     5     4     2     2 
##    58    59    60    61    62    63    65    66    67    68    69    70    71 
##     9     3     1     1     3     2     2     4     2     8     3     3     2 
##    74    77    79    81    82    88    94    97   104   105   107   108   110 
##     1     1     2     1     4     1     2     3     2     1     1     3     1 
##   111   115   117   121   124   126   128   129   138   141   143   144   146 
##     1     1     1     1     2     3     2     1     6     3     1     1     2 
##   169   171   176   179   299 
##     1     1     2     1     1

hist(data$pc_tiem_1er_prod_abierto_total)

names(data)[18] # pc_cant_moras_60_ult_12_meses: Cantidad de moras 60 en los últimos 12 meses <= 1 o vacío.

## [1] "pc_cant_moras_60_ult_12_meses"

table(data$pc_cant_moras_60_ult_12_meses)

## 
##     0     1     2     3     4     5     6     7     8     9    10    11    12 
## 14905  1052   528   353   230   130   147   175   128   132   105    87    74

hist(data$pc_cant_moras_60_ult_12_meses) # se optará por hacerla binaria

x <- ifelse(data$pc_cant_moras_60_ult_12_meses >= 1, 1, 0)
table(x)

## x
##     0     1 
## 14905  3141

flag_mora60_ult12meses <- x

names(data)[19] # gestiones_eficaces: Cantidad de gestiones eficaces

## [1] "gestiones_eficaces"

table(data$gestiones_eficaces)

## 
##     0     1     2     3 
## 18031    11     3     1

names(data)[20] # pc_transaccional: Ingreso de acuerdo al estimador transaccional del cliente

## [1] "pc_transaccional"

summary(data$pc_transaccional)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0  2043200  7547230 14096941 19069978 92269310

boxplot(data$pc_transaccional) # valores atípicos en el lado superior de la variable

quantile(data$pc_transaccional, seq(0, 1, by = 0.01))

##         0%         1%         2%         3%         4%         5%         6% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##         7%         8%         9%        10%        11%        12%        13% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        14%        15%        16%        17%        18%        19%        20% 
##        0.0        0.0    57600.0   248650.0   410806.8   665000.0   910000.0 
##        21%        22%        23%        24%        25%        26%        27% 
##  1089945.0  1381618.2  1616750.0  1840077.7  2043200.0  2266212.4  2500000.0 
##        28%        29%        30%        31%        32%        33%        34% 
##  2667730.5  2880776.6  3000000.0  3192751.5  3356433.2  3512153.2  3723150.6 
##        35%        36%        37%        38%        39%        40%        41% 
##  3963014.5  4201883.2  4410000.0  4560285.8  4812979.0  5058932.8  5239323.0 
##        42%        43%        44%        45%        46%        47%        48% 
##  5555416.6  5822665.3  6149353.8  6386744.2  6600000.0  6810155.7  7000000.0 
##        49%        50%        51%        52%        53%        54%        55% 
##  7276333.0  7547229.5  7944952.3  8239311.7  8461116.5  8765504.8  9050000.0 
##        56%        57%        58%        59%        60%        61%        62% 
##  9400018.2  9727957.2 10067467.5 10400000.0 10780000.0 11350000.0 11863116.3 
##        63%        64%        65%        66%        67%        68%        69% 
## 12154410.1 12560835.0 12988541.8 13632351.8 14295950.0 14702743.6 15137238.4 
##        70%        71%        72%        73%        74%        75%        76% 
## 15629789.5 16275052.0 16971838.3 17545912.3 18452323.0 19069977.5 19750000.0 
##        77%        78%        79%        80%        81%        82%        83% 
## 20328585.3 21272355.0 22094848.0 23043772.4 24400000.0 25657795.3 27167290.0 
##        84%        85%        86%        87%        88%        89%        90% 
## 28450000.0 30005655.5 31122150.0 32515000.0 34360117.8 36540016.0 38620000.0 
##        91%        92%        93%        94%        95%        96%        97% 
## 41672027.0 43349004.2 45725673.9 49521221.0 52488819.3 57918192.0 63254555.0 
##        98%        99%       100% 
## 70506644.0 79003883.0 92269310.0

x <- subset(data$pc_transaccional, data$pc_transaccional <= 92269310.0)
hist(x)

pc_transaccional <- data$pc_transaccional

names(data)[21] # dh_max_dia_entradas: Ultimo dia en que recibio alguna transaccion de ingreso de dinero

## [1] "dh_max_dia_entradas"

table(data$dh_max_dia_entradas)

## 
##    0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15 
## 6166   10   31   11   17   27   29   27   46   51   53   46   83   65   97   71 
##   16   17   18   19   20   21   22   23   24   25   26   27   28   29   30   31 
##   92   79   69   74  105  155  168  201  214  183  196  331 1360  449 2911 4629

hist(data$dh_max_dia_entradas) # se volverá binaria

x <- ifelse(data$dh_max_dia_entradas <= 15, 1, 0)
table(x)

## x
##     0     1 
## 11216  6830

flag_ultima_entrada_1quincena <- x

names(data)[22] # pc_cupo_entidad: Cupo de las tarjetas de crédito en el banco

## [1] "pc_cupo_entidad"

summary(data$pc_cupo_entidad)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0        0   137908        0 36000000

hist(data$pc_cupo_entidad)

table(ifelse(data$pc_cupo_entidad >= 1, 1, 0))

## 
##     0     1 
## 17834   212

names(data)[23] # pc_cuotas_como_ppal: Cuotas pagadas como principal

## [1] "pc_cuotas_como_ppal"

summary(data$pc_cuotas_como_ppal)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0   49160       0 8156850

hist(data$pc_cuotas_como_ppal)

names(data)[24] # dh_val_entradas: Valor total de los ingresos tomados en el mes anterior

## [1] "dh_val_entradas"

summary(data$dh_val_entradas)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0   5088944  18769832  26188088 204764342

hist(data$dh_val_entradas)

hist(sqrt(data$dh_val_entradas))

dh_val_entradas <- sqrt(data$dh_val_entradas)

names(data)[25] # pc_cant_moras_90_ult_12_meses: Cantidad de moras 90 o superiores en los últimos 12 meses = 0 o vacío.

## [1] "pc_cant_moras_90_ult_12_meses"

table(data$pc_cant_moras_90_ult_12_meses)

## 
##     0     1     2     3     4     5     6     7     8     9    10    11    12 
## 16250   454   310   163   149   151   106   108   111    92    69    33    50

hist(data$pc_cant_moras_90_ult_12_meses)

hist(sqrt(data$pc_cant_moras_90_ult_12_meses))

names(data)[26] # dh_max_dia_salidas

## [1] "dh_max_dia_salidas"

# prefiero quedarme con la misma variable pero para el mes actual

names(data)[27] # pc_cant_moras_60_ult_3_meses: Cantidad de moras 60 en los últimos 3 meses = 0 o vacío.

## [1] "pc_cant_moras_60_ult_3_meses"

table(data$pc_cant_moras_60_ult_3_meses)

## 
##     0     1     2     3 
## 15713   833   443  1057

table(ifelse(data$pc_cant_moras_60_ult_3_meses >= 1, 1, 0))

## 
##     0     1 
## 15713  2333

flag_tuvo_mora60_ult3meses <- ifelse(data$pc_cant_moras_60_ult_3_meses >= 1, 1, 0)

names(data)[28] # pc_cuota_tarjeta_de_credito: Cuota de tarjeta de crédito  reportada por CIFIN

## [1] "pc_cuota_tarjeta_de_credito"

summary(data$pc_cuota_tarjeta_de_credito)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0   14910       0 5445000

hist(data$pc_cuota_tarjeta_de_credito)

table(ifelse(data$pc_cuota_tarjeta_de_credito >= 1, 1, 0))

## 
##     0     1 
## 17803   243

names(data)[29] # cp_inicial_menos_saldo: valor inicial menos el saldo en el mes anterior por producto

## [1] "cp_inicial_menos_saldo"

summary(data$cp_inicial_menos_saldo)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[30] # pc_peor_estado_act_cta_aho: Peor estado cuenta ahorro

## [1] "pc_peor_estado_act_cta_aho"

summary(data$pc_peor_estado_act_cta_aho)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
## 0.00000 0.00000 0.00000 0.04034 0.00000 3.00000

hist(data$pc_peor_estado_act_cta_aho)

names(data)[31] # dia_pago: Dia de pago de la obligacion en el mes

## [1] "dia_pago"

table(data$dia_pago)

## 
##    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15   16 
##  599  314 4072  227  238  226  258  203  210  208  278  233  277  215 5161  245 
##   17   18   19   20   21   22   23   24   25   26   27   28   29   30   31 
## 1856  221  270  200  220  283  235  206  232  250  267  274  188  333   47

x <- ifelse(data$dia_pago <= 15, 1, 0)
table(x)

## x
##     0     1 
##  5327 12719

flag_diaPago_1quincena <- x

names(data)[32] # cp_cuotas_falta: Cantidad de cuotas faltantes

## [1] "cp_cuotas_falta"

table(data$cp_cuotas_falta)

## 
##     0 
## 18046

names(data)[33] # pcons_tarjeta_de_credito: Es el producto una tarjeta de credito

## [1] "pcons_tarjeta_de_credito"

table(data$pcons_tarjeta_de_credito)

## 
##     0     1 
## 12404  5642

pcons_tarjeta_de_credito <- data$pcons_tarjeta_de_credito

names(data)[34] # pc_cifin: Ingreso del cliente de acuerdo a CIFIN

## [1] "pc_cifin"

summary(data$pc_cifin)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0        0    77941        0 23509011

hist(data$pc_cifin)

names(data)[35] # gestiones_prod: Numero de gestiones productivas

## [1] "gestiones_prod"

table(data$gestiones_prod)

## 
##     0     1     2     3 
## 18036     8     1     1

names(data)[36] # pcons_vehiculos_sufi: Es el producto un vehiculo sufi

## [1] "pcons_vehiculos_sufi"

table(data$pcons_vehiculos_sufi)

## 
##     0     1 
## 16990  1056

names(data)[37] # cluster_recod: segmento cluster

## [1] "cluster_recod"

table(data$cluster_recod)

## 
##    1    6    7    9   13   16   19 
##   99 9105 5284   49 1867 1008  634

x <- ifelse(data$cluster_recod == 6, 1, 0)
table(x)

## x
##    0    1 
## 8941 9105

flag_es_cluster_6 <- x

names(data)[38] # dh_avg_dia_retiros_d: Dia promedio del mes en el que realiza los retiros 

## [1] "dh_avg_dia_retiros_d"

hist(data$dh_avg_dia_retiros_d)

dh_avg_dia_retiros_d <- data$dh_avg_dia_retiros_d

names(data)[39] # dmi_max_egreso_diario: Maximo egreso en un dia del mes anterior 

## [1] "dmi_max_egreso_diario"

summary(data$dmi_max_egreso_diario)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0  2396128  6685783  9264177 85362328

hist(data$dmi_max_egreso_diario)

quantile(data$dmi_max_egreso_diario, seq(0,1,by = 0.01))

##         0%         1%         2%         3%         4%         5%         6% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##         7%         8%         9%        10%        11%        12%        13% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        14%        15%        16%        17%        18%        19%        20% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        21%        22%        23%        24%        25%        26%        27% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        28%        29%        30%        31%        32%        33%        34% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        35%        36%        37%        38%        39%        40%        41% 
##        0.0    15805.6   108247.0   355864.7   579439.1   791444.0  1003106.8 
##        42%        43%        44%        45%        46%        47%        48% 
##  1152466.0  1308571.8  1472964.0  1600000.0  1782669.0  1888031.4  2008000.0 
##        49%        50%        51%        52%        53%        54%        55% 
##  2170653.4  2396128.5  2579869.1  2765324.0  2936515.0  3076251.0  3276314.0 
##        56%        57%        58%        59%        60%        61%        62% 
##  3537956.0  3805458.8  4000000.0  4148169.0  4461603.0  4688264.5  4987219.0 
##        63%        64%        65%        66%        67%        68%        69% 
##  5041687.1  5342686.0  5668527.5  6001403.5  6279611.5  6622351.2  7028000.0 
##        70%        71%        72%        73%        74%        75%        76% 
##  7400503.0  7733143.0  8032000.0  8416410.3  8911504.0  9264177.0  9856840.0 
##        77%        78%        79%        80%        81%        82%        83% 
## 10040000.0 10276101.0 10696562.4 11244800.0 11856206.8 12224783.0 12825890.5 
##        84%        85%        86%        87%        88%        89%        90% 
## 13536058.6 14305838.5 15037327.9 15742756.1 16566000.0 17589063.0 18992449.0 
##        91%        92%        93%        94%        95%        96%        97% 
## 20080000.0 21146267.0 22772981.0 25016338.8 27306876.0 30278258.2 35581436.4 
##        98%        99%       100% 
## 40866473.0 52347170.7 85362328.0

x <- subset(data$dmi_max_egreso_diario, data$dmi_max_egreso_diario <= 85362328)
hist(x)

dmi_max_egreso_diario <- data$dmi_max_egreso_diario

names(data)[40] # cpc_max_proc_deuda: Máximo del porcentaje de la deuda en el mes anterior

## [1] "cpc_max_proc_deuda"

summary(data$cpc_max_proc_deuda)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
## -28.1820   0.0000   0.2023   0.4324   0.9399  13.8796

names(data)[41] # dh_cant_otros_d: Cantidad de trasacciones de salida por concepto de otros

## [1] "dh_cant_otros_d"

summary(data$dh_cant_otros_d)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   0.000   0.000   1.000   3.092   4.000  74.000

hist(data$dh_cant_otros_d)

hist(sqrt(data$dh_cant_otros_d))

dh_cant_otros_d <- sqrt(data$dh_cant_otros_d)

names(data)[42] # pc_cont_30_lt_12m_tot_sf: Cantidad mora 30 últimos 12 meses sector financiero

## [1] "pc_cont_30_lt_12m_tot_sf"

summary(data$pc_cont_30_lt_12m_tot_sf)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##  0.0000  0.0000  0.0000  0.0681  0.0000 64.0000

hist(data$pc_cont_30_lt_12m_tot_sf)

table(data$pc_cont_30_lt_12m_tot_sf)

## 
##     0     1     2     3     5     6    64 
## 17999    13     5     3     3     5    18

names(data)[43] # pc_cant_mora90_ult_12m_total

## [1] "pc_cant_mora90_ult_12m_total"

table(data$pc_cant_mora90_ult_12m_total)

## 
##     0     1     2     3     4     5     6     7     8     9    10    11    12 
## 16882   212   145   123    84   121    79   102    97    82    57    27    35

table(ifelse(data$pc_cant_mora90_ult_12m_total >= 1, 1, 0))

## 
##     0     1 
## 16882  1164

x <- ifelse(data$pc_cant_mora90_ult_12m_total >= 1, 1, 0)
flag_tuvo_mora90_ult12M <- x

names(data)[44] # dmi_max_ingreso_diario: Maximo ingreso en un dia del mes anterior 

## [1] "dmi_max_ingreso_diario"

summary(data$dmi_max_ingreso_diario)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0   2540003   7845783  10000027 103018788

hist(data$dmi_max_ingreso_diario)

quantile(data$dmi_max_ingreso_diario, seq(0, 1, by = 0.01))

##          0%          1%          2%          3%          4%          5% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##          6%          7%          8%          9%         10%         11% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         12%         13%         14%         15%         16%         17% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         18%         19%         20%         21%         22%         23% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         24%         25%         26%         27%         28%         29% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         30%         31%         32%         33%         34%         35% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         36%         37%         38%         39%         40%         41% 
##         7.2     22203.0    195000.0    490000.0    700000.0    960900.0 
##         42%         43%         44%         45%         46%         47% 
##   1092137.9   1300000.0   1500000.0   1656152.0   1900000.0   2000000.0 
##         48%         49%         50%         51%         52%         53% 
##   2100000.0   2328568.1   2540003.0   2750003.0   3000000.0   3050000.0 
##         54%         55%         56%         57%         58%         59% 
##   3341821.6   3600000.0   3950000.0   4032169.0   4300000.0   4594069.3 
##         60%         61%         62%         63%         64%         65% 
##   4950000.0   5000000.0   5193296.1   5500000.0   6000000.0   6190591.0 
##         66%         67%         68%         69%         70%         71% 
##   6525936.2   7000000.0   7262786.6   7724640.1   8000001.0   8560257.4 
##         72%         73%         74%         75%         76%         77% 
##   9000000.0   9604260.0  10000000.0  10000026.8  10340000.0  10842262.6 
##         78%         79%         80%         81%         82%         83% 
##  11175227.0  12000000.0  12437389.0  13000000.0  13758802.2  14462555.0 
##         84%         85%         86%         87%         88%         89% 
##  15200041.0  16124906.2  17541904.3  18687254.4  19862000.0  20148078.3 
##         90%         91%         92%         93%         94%         95% 
##  21350000.0  23972400.0  25872101.6  28563056.0  30227878.6  33514149.0 
##         96%         97%         98%         99%        100% 
##  38000000.0  43250075.4  50045136.0  67883640.4 103018788.0

x <- subset(data$dmi_max_ingreso_diario, data$dmi_max_ingreso_diario <= 104909227.3)
hist(x)

dmi_max_ingreso_diario <- data$dmi_max_ingreso_diario

names(data)[45] # dh_val_otros_d: Valor total de las trasacciones de salida por concepto de otros en un mes

## [1] "dh_val_otros_d"

summary(data$dh_val_otros_d)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0    7586  513468  396635 9932459

hist(data$dh_val_otros_d)

quantile(data$dh_val_otros_d, seq(0,1,by=0.01))

##         0%         1%         2%         3%         4%         5%         6% 
##       0.00       0.00       0.00       0.00       0.00       0.00       0.00 
##         7%         8%         9%        10%        11%        12%        13% 
##       0.00       0.00       0.00       0.00       0.00       0.00       0.00 
##        14%        15%        16%        17%        18%        19%        20% 
##       0.00       0.00       0.00       0.00       0.00       0.00       0.00 
##        21%        22%        23%        24%        25%        26%        27% 
##       0.00       0.00       0.00       0.00       0.00       0.00       0.00 
##        28%        29%        30%        31%        32%        33%        34% 
##       0.00       0.00       0.00       0.00       0.00       0.00       0.00 
##        35%        36%        37%        38%        39%        40%        41% 
##       0.00       0.00       0.00       0.00       0.00       0.00       0.00 
##        42%        43%        44%        45%        46%        47%        48% 
##       0.00       0.00       0.00       0.00       0.00     900.00    1919.00 
##        49%        50%        51%        52%        53%        54%        55% 
##    3222.00    7586.00   10100.00   12483.00   13271.00   13271.00   14360.50 
##        56%        57%        58%        59%        60%        61%        62% 
##   17452.60   20760.00   23736.00   29835.00   36771.00   49855.45   62593.60 
##        63%        64%        65%        66%        67%        68%        69% 
##   75757.00   90129.00  104071.00  125011.80  146179.00  175034.00  202801.25 
##        70%        71%        72%        73%        74%        75%        76% 
##  229524.50  253039.00  280656.40  311751.40  348722.00  396634.75  436151.80 
##        77%        78%        79%        80%        81%        82%        83% 
##  482810.00  528182.00  573716.00  626437.00  689522.20  759090.00  837712.00 
##        84%        85%        86%        87%        88%        89%        90% 
##  934669.60 1013325.00 1107249.00 1229253.00 1340862.20 1509363.45 1692885.00 
##        91%        92%        93%        94%        95%        96%        97% 
## 1844536.00 2006918.00 2257572.45 2517164.00 2796438.00 3337151.80 4026117.00 
##        98%        99%       100% 
## 4825491.20 6503208.90 9932459.00

x <- subset(data$dh_val_otros_d, data$dh_val_otros_d <= 9998335.80)
hist(x)

dh_val_otros_d <- data$dh_val_otros_d

names(data)[46] # pc_ingreso_final: Ingreso de final del cliente

## [1] "pc_ingreso_final"

summary(data$pc_ingreso_final)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0  2000000  7221917 13140723 17741170 92226100

hist(data$pc_ingreso_final)

pc_ingreso_final <- data$pc_ingreso_final

names(data)[47] # dh_cant_pagos_d: Cantidad de pagos de salidas tuvo en el mes anterior

## [1] "dh_cant_pagos_d"

summary(data$dh_cant_pagos_d)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   0.000   0.000   3.000   4.864   8.000  49.000

hist(data$dh_cant_pagos_d)

dh_cant_pagos_d <- data$dh_cant_pagos_d

names(data)[48] # dmi_ingreso_total_mes: Ingreso total del mes anterior

## [1] "dmi_ingreso_total_mes"

summary(data$dmi_ingreso_total_mes)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0   5088944  18769832  26188088 204764342

hist(data$dmi_ingreso_total_mes)

quantile(data$dmi_ingreso_total_mes, seq(0, 1, by = 0.01))

##           0%           1%           2%           3%           4%           5% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##           6%           7%           8%           9%          10%          11% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          12%          13%          14%          15%          16%          17% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          18%          19%          20%          21%          22%          23% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          24%          25%          26%          27%          28%          29% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          30%          31%          32%          33%          34%          35% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          36%          37%          38%          39%          40%          41% 
##         8.00     24049.25    250005.50    630000.00   1056000.00   1400006.45 
##          42%          43%          44%          45%          46%          47% 
##   1838897.00   2199726.35   2600000.00   3000000.00   3300002.00   3790873.90 
##          48%          49%          50%          51%          52%          53% 
##   4106175.00   4564100.00   5088944.00   5550008.00   6004109.00   6639751.45 
##          54%          55%          56%          57%          58%          59% 
##   7250008.40   8010000.00   8667401.40   9450000.00  10150000.00  10675000.00 
##          60%          61%          62%          63%          64%          65% 
##  11550000.00  12200000.00  13019188.80  13904977.15  14787240.80  15525970.00 
##          66%          67%          68%          69%          70%          71% 
##  16474081.20  17555134.00  18585215.00  19600704.25  20290000.00  21541384.00 
##          72%          73%          74%          75%          76%          77% 
##  22528017.00  23630000.00  25156589.10  26188088.00  27948415.80  29405868.00 
##          78%          79%          80%          81%          82%          83% 
##  30600887.80  32212302.85  34000005.00  35944963.75  37561559.20  39459092.60 
##          84%          85%          86%          87%          88%          89% 
##  41691250.00  44302738.75  46987343.00  49143482.00  51159723.00  54010666.00 
##          90%          91%          92%          93%          94%          95% 
##  57475971.50  61466784.45  65288828.60  68862425.00  74272415.00  81843676.25 
##          96%          97%          98%          99%         100% 
##  88714237.00  98130984.65 111743381.40 132321500.00 204764342.00

x <- subset(data$dmi_ingreso_total_mes, data$dmi_ingreso_total_mes <= 205163705.0)
hist(x)

dmi_ingreso_total_mes <- data$dmi_ingreso_total_mes

names(data)[49] # dh_val_pagos_d: Valor total de las trasacciones de salida por concepto de pagos otros

## [1] "dh_val_pagos_d"

summary(data$dh_val_pagos_d)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0   551214  3019832  3753468 50357360

hist(data$dh_val_pagos_d)

quantile(data$dh_val_pagos_d, seq(0, 1, by = 0.01))

##         0%         1%         2%         3%         4%         5%         6% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##         7%         8%         9%        10%        11%        12%        13% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        14%        15%        16%        17%        18%        19%        20% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        21%        22%        23%        24%        25%        26%        27% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        28%        29%        30%        31%        32%        33%        34% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        35%        36%        37%        38%        39%        40%        41% 
##        0.0        0.0        0.0        0.0     1953.0    20000.0    40345.0 
##        42%        43%        44%        45%        46%        47%        48% 
##    63499.9    95093.0   127370.0   167836.2   227337.3   298628.5   373379.0 
##        49%        50%        51%        52%        53%        54%        55% 
##   458507.0   551214.0   645925.8   749470.0   854844.9   947913.0  1051616.0 
##        56%        57%        58%        59%        60%        61%        62% 
##  1162545.0  1283067.2  1388163.9  1500013.5  1610615.0  1710256.8  1847643.0 
##        63%        64%        65%        66%        67%        68%        69% 
##  1956650.8  2070132.4  2221932.8  2344112.0  2504857.0  2665210.0  2818400.0 
##        70%        71%        72%        73%        74%        75%        76% 
##  2961302.0  3074574.8  3221620.4  3432337.0  3589801.0  3753468.5  3964881.6 
##        77%        78%        79%        80%        81%        82%        83% 
##  4183091.0  4404183.0  4666203.0  4931637.0  5219968.5  5461909.0  5688066.2 
##        84%        85%        86%        87%        88%        89%        90% 
##  6002214.6  6270027.5  6668134.8  7100857.4  7530003.8  8157880.3  8782633.0 
##        91%        92%        93%        94%        95%        96%        97% 
##  9337860.7 10308822.0 11291463.0 12556160.2 13782306.0 15175063.0 17581258.2 
##        98%        99%       100% 
## 20688616.2 27004379.6 50357360.0

x <- subset(data$dh_val_pagos_d, data$dh_val_pagos_d <= 50357360.00)
hist(x)

dh_val_pagos_d <- data$dh_val_pagos_d

names(data)[50] # pc_gasto_familiar: Valor Gasto de familiar del cliente

## [1] "pc_gasto_familiar"

summary(data$pc_gasto_familiar)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0   723816  2070132  3368379  4335488 23067328

quantile(data$pc_gasto_familiar, seq(0,1,by=0.01))

##         0%         1%         2%         3%         4%         5%         6% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##         7%         8%         9%        10%        11%        12%        13% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        14%        15%        16%        17%        18%        19%        20% 
##        0.0        0.0    52827.5   144582.5   246500.0   350000.0   430100.0 
##        21%        22%        23%        24%        25%        26%        27% 
##   521914.7   588486.8   637956.8   687452.3   723815.7   763864.1   782813.0 
##        28%        29%        30%        31%        32%        33%        34% 
##   803882.1   836711.5   861377.8   882000.0   931155.0   975490.5  1039732.4 
##        35%        36%        37%        38%        39%        40%        41% 
##  1088643.0  1146363.2  1202858.0  1262250.0  1306491.8  1363816.0  1441944.3 
##        42%        43%        44%        45%        46%        47%        48% 
##  1519899.3  1583532.9  1653079.6  1712143.6  1775832.6  1845026.6  1910025.3 
##        49%        50%        51%        52%        53%        54%        55% 
##  2009880.8  2070131.5  2113333.7  2180887.5  2241724.8  2292477.7  2340728.6 
##        56%        57%        58%        59%        60%        61%        62% 
##  2403626.6  2454449.8  2507395.5  2550000.0  2582812.1  2629610.5  2698945.9 
##        63%        64%        65%        66%        67%        68%        69% 
##  2785419.6  2861250.0  2982375.6  3111782.2  3124333.0  3230361.3  3359112.7 
##        70%        71%        72%        73%        74%        75%        76% 
##  3545186.2  3722838.1  3892993.9  4043832.1  4196875.0  4335487.7  4530732.8 
##        77%        78%        79%        80%        81%        82%        83% 
##  4670284.2  4871802.6  5111082.5  5362225.0  5630442.3  5912383.7  6253169.6 
##        84%        85%        86%        87%        88%        89%        90% 
##  6542500.0  6842430.6  7286724.5  7673500.0  8172391.9  8594050.8  9055328.7 
##        91%        92%        93%        94%        95%        96%        97% 
##  9534963.8 10206885.0 10737379.9 11311996.9 12263970.4 13361803.4 14807103.4 
##        98%        99%       100% 
## 16718602.0 18943025.0 23067327.5

x <- subset(data$pc_gasto_familiar, data$pc_gasto_familiar <= 26582180.19)
hist(x)

pc_gasto_familiar <- data$pc_gasto_familiar

names(data)[51] # pc_cuotas_pagadas: Valor Cuotas de pagadas del cliente

## [1] "pc_cuotas_pagadas"

summary(data$pc_cuotas_pagadas)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0  290000  704829  900899 8156850

hist(data$pc_cuotas_pagadas)

quantile(data$pc_cuotas_pagadas, seq(0, 1, by = 0.01))

##        0%        1%        2%        3%        4%        5%        6%        7% 
##       0.0       0.0       0.0       0.0       0.0       0.0       0.0       0.0 
##        8%        9%       10%       11%       12%       13%       14%       15% 
##       0.0       0.0       0.0       0.0       0.0       0.0       0.0       0.0 
##       16%       17%       18%       19%       20%       21%       22%       23% 
##       0.0       0.0       0.0       0.0       0.0       0.0       0.0       0.0 
##       24%       25%       26%       27%       28%       29%       30%       31% 
##       0.0       0.0       0.0       0.0       0.0       0.0       0.0       0.0 
##       32%       33%       34%       35%       36%       37%       38%       39% 
##       0.0       0.0   43500.0   74571.4   95700.0  116000.0  118769.8  145000.0 
##       40%       41%       42%       43%       44%       45%       46%       47% 
##  150348.9  174000.0  184800.0  203700.0  216000.0  232000.0  240000.0  258000.0 
##       48%       49%       50%       51%       52%       53%       54%       55% 
##  284200.0  290000.0  290000.0  297478.3  312788.1  334276.0  346446.0  370968.9 
##       56%       57%       58%       59%       60%       61%       62%       63% 
##  392000.0  400305.1  410880.6  438368.6  460600.0  486710.1  490000.0  507349.8 
##       64%       65%       66%       67%       68%       69%       70%       71% 
##  528942.6  559000.0  602000.0  635938.0  652435.4  680000.0  727600.0  748420.0 
##       72%       73%       74%       75%       76%       77%       78%       79% 
##  789688.2  836217.2  860000.0  900898.7  935279.0  981092.5 1010402.6 1045573.2 
##       80%       81%       82%       83%       84%       85%       86%       87% 
## 1100000.0 1169898.4 1220208.2 1306574.5 1380618.1 1459464.8 1560829.4 1625384.1 
##       88%       89%       90%       91%       92%       93%       94%       95% 
## 1700700.2 1769856.3 1841142.1 2033524.0 2200101.2 2360000.0 2570558.7 2894792.0 
##       96%       97%       98%       99%      100% 
## 3178855.5 3683356.2 4743779.9 6010000.0 8156850.0

x <- subset(data$pc_cuotas_pagadas, data$pc_cuotas_pagadas <= 8156850.0)
hist(x)

pc_cuotas_pagadas <- data$pc_cuotas_pagadas

names(data)[52] # cpc_avg_proc_deuda: Promedio del porcentaje de las deudas de un cliente en el mes anterior (Porcentaje : Razón entre el saldo y el valor incial)

## [1] "cpc_avg_proc_deuda"

summary(data$cpc_avg_proc_deuda)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
## -119.1284    0.0000    0.1820    0.3917    0.8733   13.8796

hist(data$cpc_avg_proc_deuda)

quantile(data$cpc_avg_proc_deuda, seq(0,1,by=0.01))

##            0%            1%            2%            3%            4% 
## -119.12840000    0.00000000    0.00000000    0.00000000    0.00000000 
##            5%            6%            7%            8%            9% 
##    0.00000000    0.00000000    0.00000000    0.00000000    0.00000000 
##           10%           11%           12%           13%           14% 
##    0.00000000    0.00000000    0.00000000    0.00000000    0.00000000 
##           15%           16%           17%           18%           19% 
##    0.00000000    0.00000000    0.00000000    0.00000000    0.00000000 
##           20%           21%           22%           23%           24% 
##    0.00000000    0.00000000    0.00000000    0.00000000    0.00000000 
##           25%           26%           27%           28%           29% 
##    0.00000000    0.00000000    0.00000000    0.00000000    0.00000000 
##           30%           31%           32%           33%           34% 
##    0.00000000    0.00000000    0.00000000    0.00000000    0.00000000 
##           35%           36%           37%           38%           39% 
##    0.00000000    0.00000000    0.00000000    0.00000000    0.00000000 
##           40%           41%           42%           43%           44% 
##    0.00000000    0.00000000    0.00000000    0.00000000    0.00000000 
##           45%           46%           47%           48%           49% 
##    0.00317328    0.01566387    0.05337072    0.09155538    0.14000205 
##           50%           51%           52%           53%           54% 
##    0.18205318    0.22012374    0.25548425    0.28493140    0.32651283 
##           55%           56%           57%           58%           59% 
##    0.37242599    0.40897600    0.43696430    0.46846924    0.50734022 
##           60%           61%           62%           63%           64% 
##    0.53930356    0.56982057    0.59705567    0.63020920    0.65465374 
##           65%           66%           67%           68%           69% 
##    0.68039150    0.70248708    0.72589106    0.74845785    0.77251605 
##           70%           71%           72%           73%           74% 
##    0.79321046    0.81041248    0.82971103    0.84219350    0.85846768 
##           75%           76%           77%           78%           79% 
##    0.87328461    0.88886939    0.90253529    0.91428714    0.92869600 
##           80%           81%           82%           83%           84% 
##    0.94124096    0.95355144    0.96537928    0.97453072    0.98651367 
##           85%           86%           87%           88%           89% 
##    1.00066241    1.01072154    1.02341465    1.04034223    1.05742733 
##           90%           91%           92%           93%           94% 
##    1.07356220    1.08732512    1.09956045    1.11127967    1.12717757 
##           95%           96%           97%           98%           99% 
##    1.14653787    1.16522691    1.18974892    1.20539930    1.22587058 
##          100% 
##   13.87960000

names(data)[53] # cpc_sum_proc_deuda: Suma del porcentaje de las deudas de un cliente en el mes anterior (Porcentaje : Razón entre el saldo y el valor incial)

## [1] "cpc_sum_proc_deuda"

summary(data$cpc_sum_proc_deuda)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
## -119.1284    0.0000    0.1820    0.3917    0.8733   13.8796

hist(data$cpc_sum_proc_deuda)

names(data)[54] # dc_porc_prod_sin_mora: Porcentaje de productos sin mora en todo el sistema

## [1] "dc_porc_prod_sin_mora"

summary(data$dc_porc_prod_sin_mora)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
## 0.00000 0.00000 0.00000 0.01288 0.00000 1.00000

hist(data$dc_porc_prod_sin_mora)

quantile(data$dc_porc_prod_sin_mora, seq(0,1,by=0.01))

##   0%   1%   2%   3%   4%   5%   6%   7%   8%   9%  10%  11%  12%  13%  14%  15% 
##    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0 
##  16%  17%  18%  19%  20%  21%  22%  23%  24%  25%  26%  27%  28%  29%  30%  31% 
##    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0 
##  32%  33%  34%  35%  36%  37%  38%  39%  40%  41%  42%  43%  44%  45%  46%  47% 
##    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0 
##  48%  49%  50%  51%  52%  53%  54%  55%  56%  57%  58%  59%  60%  61%  62%  63% 
##    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0 
##  64%  65%  66%  67%  68%  69%  70%  71%  72%  73%  74%  75%  76%  77%  78%  79% 
##    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0 
##  80%  81%  82%  83%  84%  85%  86%  87%  88%  89%  90%  91%  92%  93%  94%  95% 
##    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0 
##  96%  97%  98%  99% 100% 
##    0    0    0    1    1

names(data)[55] # pc_ingreso_rutina_con_techo: Ingreso por rutina una vez aplicada los techos por segmento

## [1] "pc_ingreso_rutina_con_techo"

summary(data$pc_ingreso_rutina_con_techo)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0  2122000  7576047 14120244 19076750 92269310

hist(data$pc_ingreso_rutina_con_techo)

pc_ingreso_rutina_con_techo <- data$pc_ingreso_rutina_con_techo

names(data)[56] # pc_saldo_prom3_tdc_entidad: Saldo promedio de los últimos 3 meses de tarjeta de crédito en el banco

## [1] "pc_saldo_prom3_tdc_entidad"

summary(data$pc_saldo_prom3_tdc_entidad)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0        0    92238        0 30110000

hist(data$pc_saldo_prom3_tdc_entidad)

names(data)[57] # dh_cant_salidas: Cantidad de trasacciones de salida de dinero en un mes

## [1] "dh_cant_salidas"

summary(data$dh_cant_salidas)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    0.00    0.00   15.00   27.59   42.00  404.00

hist(data$dh_cant_salidas)

dh_cant_salidas <- data$dh_cant_salidas

names(data)[58] # dh_min_dia_pagos_d: Primer dia en el que realizó algún credito en el mes anterior

## [1] "dh_min_dia_pagos_d"

summary(data$dh_min_dia_pagos_d)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   0.000   0.000   2.000   3.952   5.000  31.000

hist(data$dh_min_dia_pagos_d)

dh_min_dia_pagos_d <- data$dh_min_dia_pagos_d

names(data)[59] # pc_ingreso_por_rutina: Ingreso por rutina

## [1] "pc_ingreso_por_rutina"

summary(data$pc_ingreso_por_rutina)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0  2122000  7576047 14120244 19076750 92269310

hist(data$pc_ingreso_por_rutina)

pc_ingreso_por_rutina <- data$pc_ingreso_por_rutina

names(data)[60] # dh_min_dia_pago_tarj_d: Primer dia en el que pago la tarjeta de credito en el mes anterior

## [1] "dh_min_dia_pago_tarj_d"

hist(data$dh_min_dia_pago_tarj_d)

names(data)[61] # cp_nro_cuota: Numero de cuota pactadas por producto

## [1] "cp_nro_cuota"

summary(data$cp_nro_cuota)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[62] # dmi_egreso_total_mes: Egreso total del mes anterior

## [1] "dmi_egreso_total_mes"

summary(data$dmi_egreso_total_mes)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0   5735566  19552715  28140390 204098909

hist(data$dmi_egreso_total_mes)

quantile(data$dmi_egreso_total_mes, seq(0,1,by=0.01))

##          0%          1%          2%          3%          4%          5% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##          6%          7%          8%          9%         10%         11% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         12%         13%         14%         15%         16%         17% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         18%         19%         20%         21%         22%         23% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         24%         25%         26%         27%         28%         29% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         30%         31%         32%         33%         34%         35% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         36%         37%         38%         39%         40%         41% 
##     20098.2    150464.5    469809.0   1007389.1   1364407.0   1757330.9 
##         42%         43%         44%         45%         46%         47% 
##   2139132.8   2597761.0   3055581.0   3491431.8   3875658.7   4341015.8 
##         48%         49%         50%         51%         52%         53% 
##   4728541.2   5246081.5   5735566.0   6222528.9   6750806.4   7463828.4 
##         54%         55%         56%         57%         58%         59% 
##   8156806.7   8903706.2   9676725.2  10353244.3  10949690.0  11653537.0 
##         60%         61%         62%         63%         64%         65% 
##  12507194.0  13308026.2  14158401.7  14949690.7  15668673.0  16852909.0 
##         66%         67%         68%         69%         70%         71% 
##  17823161.3  18701866.2  19649493.0  20633345.1  21568478.0  22936048.2 
##         72%         73%         74%         75%         76%         77% 
##  24164720.0  25304132.1  26608243.8  28140390.0  29543323.8  30673503.0 
##         78%         79%         80%         81%         82%         83% 
##  32490727.9  34304664.0  36063171.0  37668819.0  39804887.1  41664563.1 
##         84%         85%         86%         87%         88%         89% 
##  43936740.2  46612353.5  48890887.4  50950735.5  52832299.8  55616061.5 
##         90%         91%         92%         93%         94%         95% 
##  58941776.5  62642842.1  66264618.4  70686064.8  76333370.2  82037209.0 
##         96%         97%         98%         99%        100% 
##  90014943.2  99504343.8 113054665.2 132338269.6 204098909.0

x <- subset(data$dmi_egreso_total_mes, data$dmi_egreso_total_mes <= 209829245.0)
hist(x)

dmi_egreso_total_mes <- data$dmi_egreso_total_mes

names(data)[63] # cp_valor_inicial: valor inicial de la obligacion por producto

## [1] "cp_valor_inicial"

summary(data$cp_valor_inicial)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[64] # dh_max_dia_otros_d: Ultimo dia en el que realizó algún débito clasificado como "otros" en el mes anterior

## [1] "dh_max_dia_otros_d"

hist(data$dh_max_dia_otros_d)

x <- ifelse(data$dh_max_dia_otros_d <= 15, 1, 0)
table(x)

## x
##     0     1 
##  7892 10154

flax_maxDiaOtrosD_1quincena <- x

names(data)[65] # p_cuota_sobre_saldo: Valor de la cuota sobre saldo por producto

## [1] "cp_cuota_sobre_saldo"

summary(data$cp_cuota_sobre_saldo)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[66] # dc_sum_valor_inicial: Suma valores iniciales de obligaciones en todo el sistema financiero

## [1] "dc_sum_valor_inicial"

summary(data$dc_sum_valor_inicial)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0         0    602639         0 498786000

hist(data$dc_sum_valor_inicial)

names(data)[67] # dh_cant_pago_tarj_d: Cantidad de trasacciones de salida por concepto de pago de tarjeta de credito

## [1] "dh_cant_pago_tarj_d"

summary(data$dh_cant_pago_tarj_d)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##  0.0000  0.0000  0.0000  0.8238  1.0000 33.0000

hist(data$dh_cant_pago_tarj_d)

names(data)[68] # dh_max_dia_pagos_d: Ultimo dia en el que realizó algún credito en el mes anterior

## [1] "dh_max_dia_pagos_d"

hist(data$dh_max_dia_pagos_d)

x <- ifelse(data$dh_max_dia_pagos_d <= 15, 1, 0)
table(x)

## x
##    0    1 
## 9762 8284

flag_maxDiaPagos_d_1quincena <- x

names(data)[69] # cp_saldo_sobre_inicial: valor del saldo sobre inicial por producto mes anterior

## [1] "cp_saldo_sobre_inicial"

summary(data$cp_saldo_sobre_inicial)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[70] # pc_mediana_nom3: Mediana de los últimos 3 pagos nómina  para cálculo de retanqueo libranza

## [1] "pc_mediana_nom3"

summary(data$pc_mediana_nom3)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0        0    72055        0 64345000

hist(data$pc_mediana_nom3)

names(data)[71] # cp_esta_cuota_otro: Es el estado de la cuota otro

## [1] "cp_esta_cuota_otro"

summary(data$cp_esta_cuota_otro)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[72] # dh_max_dia_retiros_d: Ultimo dia en el que realizó reitro en el mes anterior

## [1] "dh_max_dia_retiros_d"

hist(data$dh_max_dia_retiros_d)

# variable incorrecta, el cuartil 3 es 30 y un mes tiene hasta 31 días

names(data)[73] # dh_avg_dia_entradas: Dia promedio del mes en el recibe las entradas de dinero

## [1] "dh_avg_dia_entradas"

summary(data$dh_avg_dia_entradas)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   0.000   0.000   4.161   3.892   6.125  31.000

hist(data$dh_avg_dia_entradas)

names(data)[74] # dh_avg_dia_pagos_d: Dia del mes promedio en el que hace las salidas de dinero por pagos

## [1] "dh_avg_dia_pagos_d"

x <- ifelse(data$dh_avg_dia_pagos_d <= 10, 1, 0)
table(x)

## x
##     0     1 
##    45 18001

names(data)[75] # dh_val_salidas: Valor total de las trasacciones de salida en un mes

## [1] "dh_val_salidas"

summary(data$dh_val_salidas)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0   5884591  19619223  28254748 204098909

hist(data$dh_val_salidas)

quantile(data$dh_val_salidas, seq(0,1,by=0.01))

##          0%          1%          2%          3%          4%          5% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##          6%          7%          8%          9%         10%         11% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         12%         13%         14%         15%         16%         17% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         18%         19%         20%         21%         22%         23% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         24%         25%         26%         27%         28%         29% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         30%         31%         32%         33%         34%         35% 
##         0.0         0.0         0.0         0.0     11359.0     12350.0 
##         36%         37%         38%         39%         40%         41% 
##     76400.8    342377.0    679878.3   1118683.2   1541005.0   1938851.1 
##         42%         43%         44%         45%         46%         47% 
##   2298733.0   2740303.9   3183321.0   3592226.5   4022760.0   4466078.0 
##         48%         49%         50%         51%         52%         53% 
##   4917904.6   5332810.0   5884591.0   6356795.0   6873811.8   7663007.0 
##         54%         55%         56%         57%         58%         59% 
##   8257288.2   9064578.0   9809325.0  10475107.3  11023429.0  11765300.4 
##         60%         61%         62%         63%         64%         65% 
##  12624035.0  13404973.0  14242719.6  15034490.1  15802825.6  16946976.0 
##         66%         67%         68%         69%         70%         71% 
##  17905644.0  18764929.0  19700935.4  20697842.0  21641359.5  23028731.0 
##         72%         73%         74%         75%         76%         77% 
##  24249542.8  25385925.3  26679698.7  28254747.5  29644945.6  30871852.0 
##         78%         79%         80%         81%         82%         83% 
##  32539082.0  34463859.1  36134965.0  37813640.0  39947167.0  41741354.8 
##         84%         85%         86%         87%         88%         89% 
##  43999606.0  46745102.2  48942396.4  50977400.0  52974824.8  55714561.0 
##         90%         91%         92%         93%         94%         95% 
##  59028184.5  62668751.4  66304736.6  70688799.0  76451589.0  82043497.0 
##         96%         97%         98%         99%        100% 
##  90022669.6  99748689.3 113054665.2 132338269.6 204098909.0

x <- subset(data$dh_val_salidas, data$dh_val_salidas <= 217011100)
hist(x)

dh_val_salidas <- data$dh_val_salidas

names(data)[76] # dc_sum_valor_cuota: Valor total de las trasacciones de salida en un mes

## [1] "dc_sum_valor_cuota"

summary(data$dc_sum_valor_cuota)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0        0    26756        0 10504000

hist(data$dc_sum_valor_cuota)

names(data)[77] # dh_min_dia_tras_d: Primer dia en el que realizó pago de traslado en el mes anterior

## [1] "dh_min_dia_tras_d"

hist(data$dh_min_dia_tras_d)

x <- ifelse(data$dh_min_dia_tras_d <= 10, 1, 0)
table(x)

## x
##     0     1 
##  1640 16406

names(data)[78] # cp_porc_valorcuot_ing: Relación entre el valor de la cuota sobre los ingresos por producto

## [1] "cp_porc_valorcuot_ing"

summary(data$cp_porc_valorcuot_ing)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[79] # pc_ind_ajustado: Ingreso neto disponible  del cliente ajustado

## [1] "pc_ind_ajustado"

summary(data$pc_ind_ajustado)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
## -6478753   319157  3173260  6474900  8958133 59465796

hist(data$pc_ind_ajustado)

quantile(data$pc_ind_ajustado, seq(0,1,by=0.01))

##          0%          1%          2%          3%          4%          5% 
## -6478752.60 -1584635.43  -736622.04  -363284.48  -136329.26   -28157.31 
##          6%          7%          8%          9%         10%         11% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         12%         13%         14%         15%         16%         17% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         18%         19%         20%         21%         22%         23% 
##        0.00        0.00        0.00    10854.35    95242.50   166275.00 
##         24%         25%         26%         27%         28%         29% 
##   229297.05   319157.46   414630.00   515272.66   656092.98   773018.33 
##         30%         31%         32%         33%         34%         35% 
##   851723.71   955501.35  1106847.00  1197787.50  1316281.58  1374374.25 
##         36%         37%         38%         39%         40%         41% 
##  1463705.35  1558277.54  1644887.26  1739666.44  1860434.10  1979169.46 
##         42%         43%         44%         45%         46%         47% 
##  2082131.05  2210792.04  2333189.63  2474835.75  2578853.57  2661750.00 
##         48%         49%         50%         51%         52%         53% 
##  2819938.87  2966919.38  3173260.41  3314359.51  3431853.07  3568431.40 
##         54%         55%         56%         57%         58%         59% 
##  3786186.59  4016250.00  4159682.15  4360409.42  4488738.16  4630921.05 
##         60%         61%         62%         63%         64%         65% 
##  4760138.69  4974054.87  5297034.28  5504701.54  5846234.27  6117532.40 
##         66%         67%         68%         69%         70%         71% 
##  6380682.46  6673624.40  6940216.30  7237858.50  7635737.48  7844523.75 
##         72%         73%         74%         75%         76%         77% 
##  8066551.78  8352009.55  8678789.30  8958132.53  9306000.90  9765300.34 
##         78%         79%         80%         81%         82%         83% 
## 10257365.98 10770282.61 11156934.11 11599371.56 12020464.91 12773224.12 
##         84%         85%         86%         87%         88%         89% 
## 13320885.71 13937072.75 14681655.47 15506136.16 16123060.47 17059105.95 
##         90%         91%         92%         93%         94%         95% 
## 17978599.46 19117837.97 20277254.42 22345186.29 24072384.41 25778069.55 
##         96%         97%         98%         99%        100% 
## 28130216.60 30465696.15 33894017.04 38331857.83 59465796.00

x <- subset(data$pc_ind_ajustado, data$pc_ind_ajustado <= 68107989.8)
hist(x)

names(data)[80] # dh_val_retiros_d: Valor total de las trasacciones de salida por concepto de retiros en un mes

## [1] "dh_val_retiros_d"

summary(data$dh_val_retiros_d)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0   1235198  11122177  13014757 197776495

hist(data$dh_val_retiros_d)

names(data)[81] # pc_tiem_lt_prod_abie_total: Tiempo último producto abierto

## [1] "pc_tiem_lt_prod_abie_total"

summary(data$pc_tiem_lt_prod_abie_total)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##  0.0000  0.0000  0.0000  0.0778  0.0000 58.0000

hist(data$pc_tiem_lt_prod_abie_total)

names(data)[82] # marca_info_cifin_decode: Marca Cifin(Consultado, no consultado, no encontrado, etc)

## [1] "marca_info_cifin_decode"

table(data$marca_info_cifin_decode)

## 
##     0     1     2 
## 12258   252  5536

x <- ifelse(data$marca_info_cifin_decode == 0, 1, 0)
table(x)

## x
##     0     1 
##  5788 12258

flag_encontrado_cifin <- x

names(data)[83] # dh_max_dia_pago_tarj_d: Ultimo dia en el que pago la tarjeta de credito en el mes anterior

## [1] "dh_max_dia_pago_tarj_d"

hist(data$dh_max_dia_pago_tarj_d)

names(data)[84] # dc_valobli_ing: Suma del valor inicial de las obligaciones en el sistema financiero sobre los ingresos

## [1] "dc_valobli_ing"

summary(data$dc_valobli_ing)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##  0.00000  0.00000  0.00000  0.05584  0.00000 58.56358

hist(data$dc_valobli_ing)

names(data)[85] # pc_cantidad_tdc_entidad: Cantidad tarjetas de crédito en el banco

## [1] "pc_cantidad_tdc_entidad"

table(data$pc_cantidad_tdc_entidad)

## 
##     0     1     2     3     4 
## 17833    93    56    13    51

names(data)[86] # dh_min_dia_otros_d: Primer dia en el que realizó algún débito clasificado como "otros" en el mes anterior

## [1] "dh_min_dia_otros_d"

hist(data$dh_min_dia_otros_d)

x <- ifelse(data$dh_min_dia_otros_d <= 15, 1, 0)
table(x)

## x
##     0     1 
##  1976 16070

names(data)[87] # dc_cant_obligaciones: Cantidad de obligaciones

## [1] "dc_cant_obligaciones"

table(data$dc_cant_obligaciones)

## 
##     0     1     2     3     4     5     6     7     8     9    10    11    13 
## 17809    18    10    32    22    23    23    10    62     8     2    11     1 
##    14    15    17    18    20 
##     1     4     3     6     1

summary(data$dc_cant_obligaciones)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##  0.00000  0.00000  0.00000  0.08423  0.00000 20.00000

hist(data$dc_cant_obligaciones)

x <- ifelse(data$dc_cant_obligaciones >= 1, 1, 0)
table(x)

## x
##     0     1 
## 17809   237

names(data)[88] # cpc_sum_nro_cuota: Suma de los número de cuotas de todas las obligaciones del cliente

## [1] "cpc_sum_nro_cuota"

summary(data$cpc_sum_nro_cuota)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    0.00    0.00   20.00   78.43  107.00 1108.00

names(data)[89] # cpc_avg_nro_cuota: Promedio del número de cuotas entre todos los productos del cliente

## [1] "cpc_avg_nro_cuota"

summary(data$cpc_avg_nro_cuota)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    0.00    0.00   14.00   34.64   51.00  446.00

hist(data$cpc_avg_nro_cuota)

cpc_avg_nro_cuota <- data$cpc_avg_nro_cuota

names(data)[90] # cpc_max_nro_cuota: Número de cuotas máximo de un cliente entre todos sus productos

## [1] "cpc_max_nro_cuota"

names(data)[91] # cp_saldo: Valor del saldo del mes anterior por producto

## [1] "cp_saldo"

summary(data$cp_saldo)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[92] # cp_cuota_sobre_inicial: valor de la cuota sobre el valor desembolsado incial por producto

## [1] "cp_cuota_sobre_inicial"

summary(data$cp_cuota_sobre_inicial)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[93] # cpc_sum_saldo: Suma del saldo de todos los productos activos del cliente

## [1] "cpc_sum_saldo"

summary(data$cpc_sum_saldo)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##  -1129268         0   1108448  11329613   8989011 220890667

hist(data$cpc_sum_saldo)

quantile(data$cpc_sum_saldo, seq(0,1,by=0.01))

##          0%          1%          2%          3%          4%          5% 
##  -1129268.0         0.0         0.0         0.0         0.0         0.0 
##          6%          7%          8%          9%         10%         11% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         12%         13%         14%         15%         16%         17% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         18%         19%         20%         21%         22%         23% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         24%         25%         26%         27%         28%         29% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         30%         31%         32%         33%         34%         35% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         36%         37%         38%         39%         40%         41% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         42%         43%         44%         45%         46%         47% 
##         0.0         0.0         0.0     33950.0    110565.4    348394.0 
##         48%         49%         50%         51%         52%         53% 
##    613901.0    900660.1   1108447.5   1439348.3   1743853.2   2009279.1 
##         54%         55%         56%         57%         58%         59% 
##   2230500.0   2464656.0   2738263.0   3022119.5   3254198.7   3529597.0 
##         60%         61%         62%         63%         64%         65% 
##   3829456.0   4079343.0   4345954.0   4611392.3   4843450.0   5057907.0 
##         66%         67%         68%         69%         70%         71% 
##   5295887.0   5527366.0   5700842.4   5994608.2   6311746.0   6793297.0 
##         72%         73%         74%         75%         76%         77% 
##   7411579.2   7832990.0   8379714.0   8989011.2   9577423.2  10105086.2 
##         78%         79%         80%         81%         82%         83% 
##  10746466.0  11413477.4  12158710.0  12852364.0  14071157.1  14697477.0 
##         84%         85%         86%         87%         88%         89% 
##  15465740.8  16480607.0  17754309.9  19526399.3  21521641.0  23601753.0 
##         90%         91%         92%         93%         94%         95% 
##  27460463.5  32852491.0  39550370.6  45636780.7  56445784.0  68892210.8 
##         96%         97%         98%         99%        100% 
##  79424456.4  96412554.3 116216129.4 145505224.0 220890667.4

x <- subset(data$cpc_sum_saldo, data$cpc_sum_saldo <= 220905770.00)
hist(x)

cpc_sum_saldo <- data$cpc_sum_saldo

names(data)[94] # cp_porc_saldo_ing: valor del saldo sobre ingreso por producto

## [1] "cp_porc_saldo_ing"

summary(data$cp_porc_saldo_ing)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[95] # gsm_mejor_gestion: Mejor gestión realizada en el mes anterior

## [1] "gsm_mejor_gestion"

table(data$gsm_mejor_gestion)

## 
##     0    16 
## 18044     2

names(data)[96] # dh_min_dia_nomina_c

## [1] "dh_min_dia_nomina_c"

names(data)[97] # dh_max_dia_nomina_c: Ultimo dia en el que recibió pago de nomina en el mes anterior

## [1] "dh_max_dia_nomina_c"

hist(data$dh_max_dia_nomina_c)

names(data)[98] # cp_valor_cuota: Valor de cuota por producto

## [1] "cp_valor_cuota"

summary(data$cp_valor_cuota)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[99] # cpc_nro_cuota_tdc: Suma de los número de cuotas de todas las tarjetas de crédito del cliente

## [1] "cpc_nro_cuota_tdc"

names(data)[100] # gsm_prom_dias_gest: Promedio de los días en que se realizaron gestiones en el mes anterior

## [1] "gsm_prom_dias_gest"

names(data)[101] # pc_cuota_no_rot_ent: Cuota de productos no rotativos en el banco

## [1] "pc_cuota_no_rot_ent"

summary(data$pc_cuota_no_rot_ent)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0   23699       0 6342000

hist(data$pc_cuota_no_rot_ent)

names(data)[102] # dh_val_nomina_c: Valor total de las trasacciones de entrada por concepto de nomina en un mes

## [1] "dh_val_nomina_c"

summary(data$dh_val_nomina_c)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0        0   101166        0 42000000

hist(data$dh_val_nomina_c)

quantile(data$dh_val_nomina_c, seq(0,1,by=0.01))

##       0%       1%       2%       3%       4%       5%       6%       7% 
##        0        0        0        0        0        0        0        0 
##       8%       9%      10%      11%      12%      13%      14%      15% 
##        0        0        0        0        0        0        0        0 
##      16%      17%      18%      19%      20%      21%      22%      23% 
##        0        0        0        0        0        0        0        0 
##      24%      25%      26%      27%      28%      29%      30%      31% 
##        0        0        0        0        0        0        0        0 
##      32%      33%      34%      35%      36%      37%      38%      39% 
##        0        0        0        0        0        0        0        0 
##      40%      41%      42%      43%      44%      45%      46%      47% 
##        0        0        0        0        0        0        0        0 
##      48%      49%      50%      51%      52%      53%      54%      55% 
##        0        0        0        0        0        0        0        0 
##      56%      57%      58%      59%      60%      61%      62%      63% 
##        0        0        0        0        0        0        0        0 
##      64%      65%      66%      67%      68%      69%      70%      71% 
##        0        0        0        0        0        0        0        0 
##      72%      73%      74%      75%      76%      77%      78%      79% 
##        0        0        0        0        0        0        0        0 
##      80%      81%      82%      83%      84%      85%      86%      87% 
##        0        0        0        0        0        0        0        0 
##      88%      89%      90%      91%      92%      93%      94%      95% 
##        0        0        0        0        0        0        0        0 
##      96%      97%      98%      99%     100% 
##        0        0   600000  2173865 42000000

names(data)[103] # banca_completa: Si el cliente pertenece al segmento banca completa o no

## [1] "banca_completa"

table(data$banca_completa)

## 
##     0 
## 18046

names(data)[104] # cpc_saldo_sobre_ing: saldo sobre ingreso por cliente

## [1] "cpc_saldo_sobre_ing"

summary(data$cpc_saldo_sobre_ing)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
## -0.07529  0.00000  0.03420  0.70105  0.50141 20.23908

hist(data$cpc_saldo_sobre_ing)

quantile(data$cpc_saldo_sobre_ing, seq(0,1,by=0.01))

##           0%           1%           2%           3%           4%           5% 
## -0.075284533  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000 
##           6%           7%           8%           9%          10%          11% 
##  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000 
##          12%          13%          14%          15%          16%          17% 
##  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000 
##          18%          19%          20%          21%          22%          23% 
##  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000 
##          24%          25%          26%          27%          28%          29% 
##  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000 
##          30%          31%          32%          33%          34%          35% 
##  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000 
##          36%          37%          38%          39%          40%          41% 
##  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000  0.000000000 
##          42%          43%          44%          45%          46%          47% 
##  0.000000000  0.000000000  0.000000000  0.001327638  0.003920323  0.010485818 
##          48%          49%          50%          51%          52%          53% 
##  0.017391611  0.026439748  0.034204096  0.042727976  0.053022024  0.064025658 
##          54%          55%          56%          57%          58%          59% 
##  0.072661546  0.082753374  0.095978925  0.109599685  0.119360811  0.133522179 
##          60%          61%          62%          63%          64%          65% 
##  0.145511888  0.153764215  0.165604222  0.180229507  0.195539646  0.214586774 
##          66%          67%          68%          69%          70%          71% 
##  0.239741468  0.260049473  0.280311386  0.297297228  0.321735771  0.349650555 
##          72%          73%          74%          75%          76%          77% 
##  0.378458794  0.413185699  0.459745553  0.501406376  0.534302699  0.570659962 
##          78%          79%          80%          81%          82%          83% 
##  0.614790560  0.670961859  0.723906664  0.810553439  0.862300980  0.949874583 
##          84%          85%          86%          87%          88%          89% 
##  1.034139144  1.126322423  1.224357805  1.353146272  1.492494873  1.647897690 
##          90%          91%          92%          93%          94%          95% 
##  1.813319044  1.978121143  2.180685275  2.429309111  2.848231115  3.195820189 
##          96%          97%          98%          99%         100% 
##  4.530609759  6.130061812  7.594766901  9.837172114 20.239077899

x <- subset(data$cpc_saldo_sobre_ing, data$cpc_saldo_sobre_ing <= 20.307294513)
hist(x)

cpc_saldo_sobre_ing <- data$cpc_saldo_sobre_ing

names(data)[105] # dh_min_dia_pago_cred_d: Primer dia en el que realizó pago de credito en el mes anterior

## [1] "dh_min_dia_pago_cred_d"

hist(data$dh_min_dia_pago_cred_d)

dh_min_dia_pago_cred_d <- data$dh_min_dia_pago_cred_d

names(data)[106] # cpc_saldo_tdc: Saldo total en tarjetas de crédito del cliente

## [1] "cpc_saldo_tdc"

summary(data$cpc_saldo_tdc)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
## -1129268        0   318800  4323266  5989105 42938793

hist(data$cpc_saldo_tdc)

quantile(data$cpc_saldo_tdc, seq(0,1,by=0.01))

##         0%         1%         2%         3%         4%         5%         6% 
## -1129268.0      -45.1        0.0        0.0        0.0        0.0        0.0 
##         7%         8%         9%        10%        11%        12%        13% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        14%        15%        16%        17%        18%        19%        20% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        21%        22%        23%        24%        25%        26%        27% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        28%        29%        30%        31%        32%        33%        34% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        35%        36%        37%        38%        39%        40%        41% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        42%        43%        44%        45%        46%        47%        48% 
##        0.0        0.0        0.0        0.0        0.0     4536.0    39499.0 
##        49%        50%        51%        52%        53%        54%        55% 
##   131904.3   318800.0   558948.6   796776.6  1028169.0  1249210.2  1555273.0 
##        56%        57%        58%        59%        60%        61%        62% 
##  1809553.0  2020438.0  2214443.8  2427540.6  2635538.0  2846335.5  3087026.0 
##        63%        64%        65%        66%        67%        68%        69% 
##  3329693.5  3592291.4  3829456.0  4031553.2  4257540.0  4491031.6  4722643.2 
##        70%        71%        72%        73%        74%        75%        76% 
##  4932761.5  5111416.0  5366764.6  5540825.0  5735475.3  5989105.0  6209213.4 
##        77%        78%        79%        80%        81%        82%        83% 
##  6643593.0  7185642.5  7609331.2  8034478.0  8497649.0  9048885.3  9573365.3 
##        84%        85%        86%        87%        88%        89%        90% 
## 10016976.0 10596398.2 11109566.0 11723089.0 12303712.2 13004694.4 14105014.5 
##        91%        92%        93%        94%        95%        96%        97% 
## 14697477.0 15541141.0 16460852.1 17506625.0 19008508.5 20708956.0 22249488.6 
##        98%        99%       100% 
## 25956592.3 30683139.0 42938793.0

x <- subset(data$cpc_saldo_tdc, data$cpc_saldo_tdc <= 43186883.96)
hist(x)

cpc_saldo_tdc <- data$cpc_saldo_tdc

names(data)[107] # pc_cuota_de_consumo: Cuota de crédito  de consumo reportada por CIFIN

## [1] "pc_cuota_de_consumo"

summary(data$pc_cuota_de_consumo)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0   22909       0 6342000

hist(data$pc_cuota_de_consumo)

names(data)[108] # dh_cant_tras_d: Cantidad de traslados de salida de dinero en un mes

## [1] "dh_cant_tras_d"

hist(data$dh_cant_tras_d)

quantile(data$dh_cant_tras_d, seq(0,1,by=0.005))

##   0.0%   0.5%   1.0%   1.5%   2.0%   2.5%   3.0%   3.5%   4.0%   4.5%   5.0% 
##      0      0      0      0      0      0      0      0      0      0      0 
##   5.5%   6.0%   6.5%   7.0%   7.5%   8.0%   8.5%   9.0%   9.5%  10.0%  10.5% 
##      0      0      0      0      0      0      0      0      0      0      0 
##  11.0%  11.5%  12.0%  12.5%  13.0%  13.5%  14.0%  14.5%  15.0%  15.5%  16.0% 
##      0      0      0      0      0      0      0      0      0      0      0 
##  16.5%  17.0%  17.5%  18.0%  18.5%  19.0%  19.5%  20.0%  20.5%  21.0%  21.5% 
##      0      0      0      0      0      0      0      0      0      0      0 
##  22.0%  22.5%  23.0%  23.5%  24.0%  24.5%  25.0%  25.5%  26.0%  26.5%  27.0% 
##      0      0      0      0      0      0      0      0      0      0      0 
##  27.5%  28.0%  28.5%  29.0%  29.5%  30.0%  30.5%  31.0%  31.5%  32.0%  32.5% 
##      0      0      0      0      0      0      0      0      0      0      0 
##  33.0%  33.5%  34.0%  34.5%  35.0%  35.5%  36.0%  36.5%  37.0%  37.5%  38.0% 
##      0      0      0      0      0      0      0      0      0      0      0 
##  38.5%  39.0%  39.5%  40.0%  40.5%  41.0%  41.5%  42.0%  42.5%  43.0%  43.5% 
##      0      0      0      0      0      0      0      0      0      0      0 
##  44.0%  44.5%  45.0%  45.5%  46.0%  46.5%  47.0%  47.5%  48.0%  48.5%  49.0% 
##      0      0      0      0      0      0      0      0      0      0      0 
##  49.5%  50.0%  50.5%  51.0%  51.5%  52.0%  52.5%  53.0%  53.5%  54.0%  54.5% 
##      0      0      0      0      0      0      0      0      0      0      0 
##  55.0%  55.5%  56.0%  56.5%  57.0%  57.5%  58.0%  58.5%  59.0%  59.5%  60.0% 
##      0      0      0      0      0      0      0      0      0      0      0 
##  60.5%  61.0%  61.5%  62.0%  62.5%  63.0%  63.5%  64.0%  64.5%  65.0%  65.5% 
##      0      0      0      0      0      0      0      0      0      0      1 
##  66.0%  66.5%  67.0%  67.5%  68.0%  68.5%  69.0%  69.5%  70.0%  70.5%  71.0% 
##      1      1      1      1      1      1      1      1      1      1      1 
##  71.5%  72.0%  72.5%  73.0%  73.5%  74.0%  74.5%  75.0%  75.5%  76.0%  76.5% 
##      1      1      2      2      2      2      2      2      2      2      2 
##  77.0%  77.5%  78.0%  78.5%  79.0%  79.5%  80.0%  80.5%  81.0%  81.5%  82.0% 
##      3      3      3      3      3      3      3      4      4      4      4 
##  82.5%  83.0%  83.5%  84.0%  84.5%  85.0%  85.5%  86.0%  86.5%  87.0%  87.5% 
##      4      4      5      5      5      5      6      6      6      6      7 
##  88.0%  88.5%  89.0%  89.5%  90.0%  90.5%  91.0%  91.5%  92.0%  92.5%  93.0% 
##      7      8      8      8      9      9      9     10     10     11     11 
##  93.5%  94.0%  94.5%  95.0%  95.5%  96.0%  96.5%  97.0%  97.5%  98.0%  98.5% 
##     12     12     13     14     14     15     16     17     18     19     20 
##  99.0%  99.5% 100.0% 
##     23     25     31

x <- subset(data$dh_cant_tras_d, data$dh_cant_tras_d <= 31)
hist(x)

dh_cant_tras_d <- data$dh_cant_tras_d

names(data)[109] # dh_max_dia_comisio_d

## [1] "dh_max_dia_comisio_d"

names(data)[110] # cpc_avg_saldo: Promedio del saldo de las obligaciones del cliente en el mes anterior

## [1] "cpc_avg_saldo"

summary(data$cpc_avg_saldo)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##  -1011469         0    774422   5152861   4793250 106882802

hist(data$cpc_avg_saldo)

quantile(data$cpc_avg_saldo, seq(0,1,by=0.01))

##          0%          1%          2%          3%          4%          5% 
##  -1011469.0         0.0         0.0         0.0         0.0         0.0 
##          6%          7%          8%          9%         10%         11% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         12%         13%         14%         15%         16%         17% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         18%         19%         20%         21%         22%         23% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         24%         25%         26%         27%         28%         29% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         30%         31%         32%         33%         34%         35% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         36%         37%         38%         39%         40%         41% 
##         0.0         0.0         0.0         0.0         0.0         0.0 
##         42%         43%         44%         45%         46%         47% 
##         0.0         0.0         0.0     16975.0     69850.0    206899.0 
##         48%         49%         50%         51%         52%         53% 
##    365970.0    565479.8    774421.5    940160.4   1053110.0   1202192.1 
##         54%         55%         56%         57%         58%         59% 
##   1358774.5   1517750.0   1668502.5   1818995.1   1946590.0   2064736.0 
##         60%         61%         62%         63%         64%         65% 
##   2199797.0   2328455.4   2429205.4   2528953.5   2676266.0   2793238.0 
##         66%         67%         68%         69%         70%         71% 
##   2973823.1   3124962.0   3305366.2   3512956.1   3720362.0   3908560.0 
##         72%         73%         74%         75%         76%         77% 
##   4146008.0   4315487.0   4534635.6   4793250.0   4927245.0   5105438.0 
##         78%         79%         80%         81%         82%         83% 
##   5327563.0   5540825.0   5755606.8   6023367.8   6305529.5   6769565.3 
##         84%         85%         86%         87%         88%         89% 
##   7375373.0   7884712.2   8498791.6   9327747.6  10567824.0  11633631.6 
##         90%         91%         92%         93%         94%         95% 
##  13229350.7  15008544.5  17404416.0  19849268.4  21885903.6  25137925.4 
##         96%         97%         98%         99%        100% 
##  33473766.1  40065113.3  47628017.1  66073881.0 106882801.8

x <- subset(data$cpc_avg_saldo, data$cpc_avg_saldo <= 106884393.50)
hist(x)

cpc_avg_saldo <- data$cpc_avg_saldo

names(data)[111] # dc_max_saldo_sf: máximo saldo en el sistema financiero (Solo Financiero)

## [1] "dc_max_saldo_sf"

summary(data$dc_max_saldo_sf)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0         0    259660         0 182310000

names(data)[112] # dh_val_pago_tarj_d: Valor total de las trasacciones de salida por concepto de pagos de tdc en un mes

## [1] "dh_val_pago_tarj_d"

summary(data$dh_val_pago_tarj_d)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0        0   427965   310071 10559486

hist(data$dh_val_pago_tarj_d)

quantile(data$dh_val_pago_tarj_d, seq(0,1,by=0.01))

##         0%         1%         2%         3%         4%         5%         6% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##         7%         8%         9%        10%        11%        12%        13% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        14%        15%        16%        17%        18%        19%        20% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        21%        22%        23%        24%        25%        26%        27% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        28%        29%        30%        31%        32%        33%        34% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        35%        36%        37%        38%        39%        40%        41% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        42%        43%        44%        45%        46%        47%        48% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        49%        50%        51%        52%        53%        54%        55% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        56%        57%        58%        59%        60%        61%        62% 
##        0.0        0.0        0.0        0.0        0.0        0.0        0.0 
##        63%        64%        65%        66%        67%        68%        69% 
##        0.0        0.0        0.0     6823.0    20000.0    34674.4    61262.1 
##        70%        71%        72%        73%        74%        75%        76% 
##   113870.0   154310.6   190467.0   230752.0   268506.0   310071.0   353957.8 
##        77%        78%        79%        80%        81%        82%        83% 
##   401450.0   440004.8   485001.7   549448.0   611841.9   681124.1   720438.2 
##        84%        85%        86%        87%        88%        89%        90% 
##   795860.0   864850.0   946432.5  1017867.0  1122199.0  1239723.7  1383143.5 
##        91%        92%        93%        94%        95%        96%        97% 
##  1511667.0  1671642.8  1883813.3  2088449.4  2368354.8  2818150.4  3187581.8 
##        98%        99%       100% 
##  3910841.7  5267074.4 10559486.0

x <- subset(data$dh_val_pago_tarj_d, data$dh_val_pago_tarj_d <= 10560315.0)
hist(x)

dh_val_pago_tarj_d <- data$dh_val_pago_tarj_d

names(data)[113] # pc_productos_no_rotativos_entidad: Cantidad de productos no rotativos en el banco

## [1] "pc_productos_no_rotativos_entidad"

table(data$pc_productos_no_rotativos_entidad) 

## 
##     0     1     2     3     4     5     6     9    18 
## 17850    54    38    20    36     4    10     1    33

hist(data$pc_productos_no_rotativos_entidad)

names(data)[114] # pc_saldo_no_rot_ent: Obligaciones no rotativos en el banco

## [1] "pc_saldo_no_rot_ent"

summary(data$pc_saldo_no_rot_ent)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0         0    918154         0 229052000

names(data)[115] # pc_vi_no_rotativos_entidad: Obligaciones no rotativos en el banco

## [1] "pc_vi_no_rotativos_entidad"

summary(data$pc_vi_no_rotativos_entidad)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0    1033       0  233279

names(data)[116] # dh_min_dia_entradas

## [1] "dh_min_dia_entradas"

names(data)[117] # pcons_hipotecario_vivienda: Es el producto un hipotecario

## [1] "pcons_hipotecario_vivienda"

table(data$pcons_hipotecario_vivienda)

## 
##     0     1 
## 17500   546

pcons_hipotecario_vivienda <- data$pcons_hipotecario_vivienda

names(data)[118] # gsm_mejor_gestion_3m: Mejor gestión realizada en los últimos 3 meses

## [1] "gsm_mejor_gestion_3m"

table(data$gsm_mejor_gestion_3m)

## 
##     0    16 
## 18044     2

names(data)[119] # dh_avg_dia_salidas: Dia promedio del mes en el que hace las salidas de dinero

## [1] "dh_avg_dia_salidas"

names(data)[120] # y_auto_cura: Variable respuesta: si el cliente se autocuro o no

## [1] "y_auto_cura"

table(data$y_auto_cura)

## 
##     0     1 
##  5590 12456

Y <- data$y_auto_cura

names(data)[122] # segmentoestructural: Segmento  estructural

## [1] "segmentoestructural"

table(data$segmentoestructural)

## 
##  PYME 
## 18046

names(data)[123] # subsegmentoestructural

## [1] "subsegmentoestructural"

table(data$subsegmentoestructural)

## 
## Pyme Pequena 
##        18046

Unificación de variables

unificado <- as.data.frame(
  cbind(
    Y,
    max_sem,
    desv_sem,
    prom_bim,
    max_mes_anterior,
    prom_mes_anterior,
    prom_sem,
    max_bim,
    prom_trim,
    pc_cant_moras_30_ult_12_meses,
    desv_trim,
    desv_bim,
    dh_cant_entradas,
    pc_transaccional,
    dh_val_entradas,
    pcons_tarjeta_de_credito,
    dh_avg_dia_retiros_d,
    dmi_max_egreso_diario,
    dh_cant_otros_d,
    dmi_max_ingreso_diario,
    dh_val_otros_d,
    pc_ingreso_final,
    dh_cant_pagos_d,
    dmi_ingreso_total_mes,
    dh_val_pagos_d,
    pc_gasto_familiar,
    pc_cuotas_pagadas,
    pc_ingreso_rutina_con_techo,
    dh_cant_salidas,
    dh_min_dia_pagos_d,
    pc_ingreso_por_rutina,
    dmi_egreso_total_mes,
    dh_val_salidas,
    cpc_avg_nro_cuota,
    cpc_sum_saldo,
    cpc_saldo_sobre_ing,
    dh_min_dia_pago_cred_d,
    cpc_saldo_tdc,
    dh_cant_tras_d,
    cpc_avg_saldo,
    dh_val_pago_tarj_d,
    pcons_hipotecario_vivienda,
    flag_mora60_ult12meses,
    flag_ultima_entrada_1quincena,
    flag_tuvo_mora60_ult3meses,
    flag_diaPago_1quincena,
    flag_es_cluster_6,
    flag_tuvo_mora90_ult12M,
    flax_maxDiaOtrosD_1quincena,
    flag_maxDiaPagos_d_1quincena,
    flag_encontrado_cifin
  )
)

dim(unificado)

## [1] 18046    51

str(unificado)

## 'data.frame':    18046 obs. of  51 variables:
##  $ Y                            : num  0 0 0 0 0 0 1 1 1 1 ...
##  $ max_sem                      : num  7 17 36 27 32 32 0 0 0 12 ...
##  $ desv_sem                     : num  2.86 7.64 12.87 10.89 11.65 ...
##  $ prom_bim                     : num  1.87 3.46 3.67 3.67 3.16 ...
##  $ max_mes_anterior             : num  7 17 27 27 20 27 0 0 0 12 ...
##  $ prom_mes_anterior            : num  2.65 4.12 5.2 5.2 4.47 ...
##  $ prom_sem                     : num  1.17 8 21 4.83 20.83 ...
##  $ max_bim                      : num  7 17 27 27 20 30 0 0 0 12 ...
##  $ prom_trim                    : num  1.53 3.7 3.74 3 4.16 ...
##  $ pc_cant_moras_30_ult_12_meses: num  1 1 0 1 0 0 0 0 0 0 ...
##  $ desv_trim                    : num  2.01 2.4 3.68 3.95 4.02 ...
##  $ desv_bim                     : num  2.22 2.66 4.37 4.37 3.76 ...
##  $ dh_cant_entradas             : num  0 3.74 2.24 0 1.41 ...
##  $ pc_transaccional             : num  3000000 7710300 6670935 3000000 3150000 ...
##  $ dh_val_entradas              : num  0 2110 1785 0 1000 ...
##  $ pcons_tarjeta_de_credito     : num  0 0 0 0 0 0 1 0 0 1 ...
##  $ dh_avg_dia_retiros_d         : num  0 16.6 22 0 16.2 ...
##  $ dmi_max_egreso_diario        : num  0 2061032 2131531 0 1002318 ...
##  $ dh_cant_otros_d              : num  0 1.41 1.41 0 0 ...
##  $ dmi_max_ingreso_diario       : num  0 2210300 1858600 0 1000000 ...
##  $ dh_val_otros_d               : num  0 22651 20200 0 0 ...
##  $ pc_ingreso_final             : num  3000000 6553755 6670935 3000000 2677500 ...
##  $ dh_cant_pagos_d              : num  0 8 1 0 4 1 0 10 13 0 ...
##  $ dmi_ingreso_total_mes        : num  0 4452485 3187600 0 1000000 ...
##  $ dh_val_pagos_d               : num  0 3085738 1 0 1978774 ...
##  $ pc_gasto_familiar            : num  765000 2313090 1701088 765000 945000 ...
##  $ pc_cuotas_pagadas            : num  0 0 0 0 0 ...
##  $ pc_ingreso_rutina_con_techo  : num  3000000 7710300 6670935 3000000 3150000 ...
##  $ dh_cant_salidas              : num  0 28 11 0 10 11 0 72 69 0 ...
##  $ dh_min_dia_pagos_d           : num  0 13 7 0 2 7 0 9 8 0 ...
##  $ pc_ingreso_por_rutina        : num  3000000 7710300 6670935 3000000 3150000 ...
##  $ dmi_egreso_total_mes         : num  0 4921015 3055581 0 1998986 ...
##  $ dh_val_salidas               : num  0 4921015 3055581 0 1998986 ...
##  $ cpc_avg_nro_cuota            : num  0 0 0 0 0 0 0 64 70 0 ...
##  $ cpc_sum_saldo                : num  0 0 0 0 0 ...
##  $ cpc_saldo_sobre_ing          : num  0 0 0 0 0 ...
##  $ dh_min_dia_pago_cred_d       : num  0 13 7 0 2 7 0 17 15 0 ...
##  $ cpc_saldo_tdc                : num  0 0 0 0 0 ...
##  $ dh_cant_tras_d               : num  0 0 0 0 0 0 0 4 2 0 ...
##  $ cpc_avg_saldo                : num  0 0 0 0 0 ...
##  $ dh_val_pago_tarj_d           : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ pcons_hipotecario_vivienda   : num  0 0 0 0 0 0 0 0 1 0 ...
##  $ flag_mora60_ult12meses       : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ flag_ultima_entrada_1quincena: num  1 0 0 1 0 0 1 0 0 1 ...
##  $ flag_tuvo_mora60_ult3meses   : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ flag_diaPago_1quincena       : num  1 1 1 0 0 0 1 0 0 0 ...
##  $ flag_es_cluster_6            : num  0 0 0 0 0 0 1 0 0 0 ...
##  $ flag_tuvo_mora90_ult12M      : num  0 0 0 0 0 0 0 0 0 0 ...
##  $ flax_maxDiaOtrosD_1quincena  : num  1 0 0 1 1 0 1 0 0 1 ...
##  $ flag_maxDiaPagos_d_1quincena : num  1 0 1 1 0 1 1 0 0 1 ...
##  $ flag_encontrado_cifin        : num  0 1 1 0 1 1 0 1 1 0 ...

Selección de variables

Se utilizará la función nearZeroVar para eliminar variables con varianza cercana a 0 y la función findCorrelation para eliminar variables independientes con correlación >= x entre sí.

Y <- unificado$Y
unificado$Y <- NULL
X <- unificado

# nearZeroVar
i <- nearZeroVar(X)
# variables a omitir por no tener varianza
names(X)[i]

## [1] "dh_cant_salidas"            "cpc_avg_nro_cuota"         
## [3] "pcons_hipotecario_vivienda"

X <- X[,-i]

# FindCorrelation
i <- findCorrelation(cor(X),cutoff = 0.8)
# variables a omitir por tener alta correlación entre sí
names(X)[i]

##  [1] "dh_val_entradas"               "dh_val_salidas"               
##  [3] "dmi_egreso_total_mes"          "dmi_ingreso_total_mes"        
##  [5] "flag_maxDiaPagos_d_1quincena"  "flag_ultima_entrada_1quincena"
##  [7] "dmi_max_egreso_diario"         "pc_gasto_familiar"            
##  [9] "prom_trim"                     "pc_ingreso_rutina_con_techo"  
## [11] "pc_transaccional"              "pc_ingreso_final"             
## [13] "prom_bim"                      "pc_cant_moras_30_ult_12_meses"
## [15] "max_bim"                       "max_sem"                      
## [17] "prom_sem"                      "flag_mora60_ult12meses"       
## [19] "prom_mes_anterior"             "desv_bim"                     
## [21] "cpc_sum_saldo"

X <- X[,-i]

# variables resultantes
salida <- cbind(Y, X)
dim(salida) 

## [1] 18046    27

# Variables que quedan
names(salida)

##  [1] "Y"                           "desv_sem"                   
##  [3] "max_mes_anterior"            "desv_trim"                  
##  [5] "dh_cant_entradas"            "pcons_tarjeta_de_credito"   
##  [7] "dh_avg_dia_retiros_d"        "dh_cant_otros_d"            
##  [9] "dmi_max_ingreso_diario"      "dh_val_otros_d"             
## [11] "dh_cant_pagos_d"             "dh_val_pagos_d"             
## [13] "pc_cuotas_pagadas"           "dh_min_dia_pagos_d"         
## [15] "pc_ingreso_por_rutina"       "cpc_saldo_sobre_ing"        
## [17] "dh_min_dia_pago_cred_d"      "cpc_saldo_tdc"              
## [19] "dh_cant_tras_d"              "cpc_avg_saldo"              
## [21] "dh_val_pago_tarj_d"          "flag_tuvo_mora60_ult3meses" 
## [23] "flag_diaPago_1quincena"      "flag_es_cluster_6"          
## [25] "flag_tuvo_mora90_ult12M"     "flax_maxDiaOtrosD_1quincena"
## [27] "flag_encontrado_cifin"

Partir datasets

set.seed(1) # semilla para hacer reproducible el proceso
S <- sample(1:nrow(salida), round(nrow(salida)*0.8,0), replace = FALSE)
train <- salida[S,]
test <- salida[-S,]
# Verificación
nrow(salida) - nrow(train) - nrow(test)

## [1] 0

Scaling

Se aplicará un scaling usando punteos Z (Z = (Xi - X) / S) a través de la función preprocess del páquete CARET. Es importante destacar que los parámetros del train dataset (media y desviación) serán usados para el scaling del test y oot dataset

# aplicar scaling de punteo Z para favorecer el aprendizaje de los modelos
Y <- train$Y
train$Y <- NULL
preprocesado <- preProcess(train, scale = TRUE, center = FALSE)
train <- predict(preprocesado, train)
train <- cbind(Y, train)
summary(train)

##        Y             desv_sem       max_mes_anterior    desv_trim       
##  Min.   :0.0000   Min.   :-0.8059   Min.   :-0.5075   Min.   :-1.19411  
##  1st Qu.:0.0000   1st Qu.:-0.6508   1st Qu.:-0.5075   1st Qu.:-1.19411  
##  Median :1.0000   Median :-0.2743   Median :-0.3711   Median :-0.01921  
##  Mean   :0.6916   Mean   : 0.0000   Mean   : 0.0000   Mean   : 0.00000  
##  3rd Qu.:1.0000   3rd Qu.: 0.3089   3rd Qu.: 0.1062   3rd Qu.: 0.66893  
##  Max.   :1.0000   Max.   :25.0106   Max.   :21.1784   Max.   : 8.93227  
##  dh_cant_entradas  pcons_tarjeta_de_credito dh_avg_dia_retiros_d
##  Min.   :-1.0603   Min.   :-0.6752          Min.   :-1.2211     
##  1st Qu.:-1.0603   1st Qu.:-0.6752          1st Qu.:-1.2211     
##  Median :-0.1007   Median :-0.6752          Median : 0.3910     
##  Mean   : 0.0000   Mean   : 0.0000          Mean   : 0.0000     
##  3rd Qu.: 0.7091   3rd Qu.: 1.4809          3rd Qu.: 0.7896     
##  Max.   : 3.9993   Max.   : 1.4809          Max.   : 3.5566     
##  dh_cant_otros_d   dmi_max_ingreso_diario dh_val_otros_d    dh_cant_pagos_d  
##  Min.   :-0.8645   Min.   :-0.5939        Min.   :-0.4234   Min.   :-0.7778  
##  1st Qu.:-0.8645   1st Qu.:-0.5939        1st Qu.:-0.4234   1st Qu.:-0.7778  
##  Median :-0.1101   Median :-0.4010        Median :-0.4165   Median :-0.2944  
##  Mean   : 0.0000   Mean   : 0.0000        Mean   : 0.0000   Mean   : 0.0000  
##  3rd Qu.: 0.6443   3rd Qu.: 0.1653        3rd Qu.:-0.0938   3rd Qu.: 0.5113  
##  Max.   : 5.5814   Max.   : 7.2274        Max.   : 7.7619   Max.   : 6.3123  
##  dh_val_pagos_d    pc_cuotas_pagadas dh_min_dia_pagos_d pc_ingreso_por_rutina
##  Min.   :-0.5468   Min.   :-0.6194   Min.   :-0.6552    Min.   :-0.8032      
##  1st Qu.:-0.5468   1st Qu.:-0.6194   1st Qu.:-0.6552    1st Qu.:-0.6827      
##  Median :-0.4491   Median :-0.3631   Median :-0.3244    Median :-0.3695      
##  Mean   : 0.0000   Mean   : 0.0000   Mean   : 0.0000    Mean   : 0.0000      
##  3rd Qu.: 0.1372   3rd Qu.: 0.1808   3rd Qu.: 0.1718    3rd Qu.: 0.2769      
##  Max.   : 8.6602   Max.   : 6.5910   Max.   : 4.4722    Max.   : 4.4212      
##  cpc_saldo_sobre_ing dh_min_dia_pago_cred_d cpc_saldo_tdc    
##  Min.   :-0.4104     Min.   :-0.6821        Min.   :-0.7940  
##  1st Qu.:-0.3699     1st Qu.:-0.6821        1st Qu.:-0.6288  
##  Median :-0.3519     Median :-0.5430        Median :-0.5864  
##  Mean   : 0.0000     Mean   : 0.0000        Mean   : 0.0000  
##  3rd Qu.:-0.1032     3rd Qu.: 0.2916        3rd Qu.: 0.2453  
##  Max.   :10.5052     Max.   : 3.6301        Max.   : 5.6517  
##  dh_cant_tras_d     cpc_avg_saldo     dh_val_pago_tarj_d
##  Min.   :-0.47777   Min.   :-0.5237   Min.   :-0.3933   
##  1st Qu.:-0.47777   1st Qu.:-0.4361   1st Qu.:-0.3933   
##  Median :-0.47777   Median :-0.3738   Median :-0.3933   
##  Mean   : 0.00000   Mean   : 0.0000   Mean   : 0.0000   
##  3rd Qu.:-0.06338   3rd Qu.:-0.0262   3rd Qu.:-0.1123   
##  Max.   : 5.94527   Max.   : 8.8241   Max.   : 9.4425   
##  flag_tuvo_mora60_ult3meses flag_diaPago_1quincena flag_es_cluster_6
##  Min.   :-0.3841            Min.   :-1.5585        Min.   :-1.0087  
##  1st Qu.:-0.3841            1st Qu.:-1.5585        1st Qu.:-1.0087  
##  Median :-0.3841            Median : 0.6416        Median : 0.9913  
##  Mean   : 0.0000            Mean   : 0.0000        Mean   : 0.0000  
##  3rd Qu.:-0.3841            3rd Qu.: 0.6416        3rd Qu.: 0.9913  
##  Max.   : 2.6035            Max.   : 0.6416        Max.   : 0.9913  
##  flag_tuvo_mora90_ult12M flax_maxDiaOtrosD_1quincena flag_encontrado_cifin
##  Min.   :-0.2625         Min.   :-1.1344             Min.   :-1.4521      
##  1st Qu.:-0.2625         1st Qu.:-1.1344             1st Qu.:-1.4521      
##  Median :-0.2625         Median : 0.8815             Median : 0.6886      
##  Mean   : 0.0000         Mean   : 0.0000             Mean   : 0.0000      
##  3rd Qu.:-0.2625         3rd Qu.: 0.8815             3rd Qu.: 0.6886      
##  Max.   : 3.8087         Max.   : 0.8815             Max.   : 0.6886

# Aplicar scaling al test dataset
# se aplicará usando las medias y desviación del train dataset
Y <- test$Y
test$Y <- NULL
test <- predict(preprocesado, test)
test <- cbind(Y, test)
summary(test)

##        Y             desv_sem         max_mes_anterior     desv_trim       
##  Min.   :0.0000   Min.   :-0.805923   Min.   :-0.50753   Min.   :-1.19411  
##  1st Qu.:0.0000   1st Qu.:-0.642950   1st Qu.:-0.50753   1st Qu.:-1.19411  
##  Median :1.0000   Median :-0.287648   Median :-0.50753   Median :-0.04873  
##  Mean   :0.6847   Mean   :-0.003183   Mean   : 0.00545   Mean   :-0.01659  
##  3rd Qu.:1.0000   3rd Qu.: 0.285270   3rd Qu.: 0.10622   3rd Qu.: 0.66176  
##  Max.   :1.0000   Max.   :16.091238   Max.   :38.70451   Max.   : 7.25586  
##  dh_cant_entradas    pcons_tarjeta_de_credito dh_avg_dia_retiros_d
##  Min.   :-1.060281   Min.   :-0.675202        Min.   :-1.221137   
##  1st Qu.:-1.060281   1st Qu.:-0.675202        1st Qu.:-1.221137   
##  Median :-0.100676   Median :-0.675202        Median : 0.403731   
##  Mean   : 0.002092   Mean   :-0.005479        Mean   : 0.006123   
##  3rd Qu.: 0.760440   3rd Qu.: 1.480935        3rd Qu.: 0.789637   
##  Max.   : 3.756894   Max.   : 1.480935        Max.   : 4.059682   
##  dh_cant_otros_d     dmi_max_ingreso_diario dh_val_otros_d   
##  Min.   :-0.864524   Min.   :-0.593912      Min.   :-0.4234  
##  1st Qu.:-0.864524   1st Qu.:-0.593912      1st Qu.:-0.4234  
##  Median :-0.110089   Median :-0.402591      Median :-0.4194  
##  Mean   : 0.003838   Mean   : 0.008742      Mean   : 0.0152  
##  3rd Qu.: 0.644345   3rd Qu.: 0.173518      3rd Qu.:-0.1013  
##  Max.   : 5.625365   Max.   : 7.227389      Max.   : 7.8260  
##  dh_cant_pagos_d    dh_val_pagos_d     pc_cuotas_pagadas  dh_min_dia_pagos_d 
##  Min.   :-0.77780   Min.   :-0.54682   Min.   :-0.61941   Min.   :-0.655150  
##  1st Qu.:-0.77780   1st Qu.:-0.54682   1st Qu.:-0.61941   1st Qu.:-0.655150  
##  Median :-0.29438   Median :-0.43672   Median :-0.35645   Median :-0.324356  
##  Mean   : 0.02972   Mean   : 0.02651   Mean   : 0.01819   Mean   :-0.007585  
##  3rd Qu.: 0.51131   3rd Qu.: 0.14503   3rd Qu.: 0.16807   3rd Qu.: 0.171836  
##  Max.   : 7.11798   Max.   : 8.64759   Max.   : 6.59100   Max.   : 4.472163  
##  pc_ingreso_por_rutina cpc_saldo_sobre_ing dh_min_dia_pago_cred_d
##  Min.   :-0.80319      Min.   :-0.40385    Min.   :-0.68214      
##  1st Qu.:-0.68304      1st Qu.:-0.36990    1st Qu.:-0.68214      
##  Median :-0.37979      Median :-0.34969    Median :-0.54303      
##  Mean   :-0.01846      Mean   : 0.03398    Mean   :-0.01309      
##  3rd Qu.: 0.27872      3rd Qu.:-0.09272    3rd Qu.: 0.29160      
##  Max.   : 4.41872      Max.   :10.50524    Max.   : 3.63014      
##  cpc_saldo_tdc      dh_cant_tras_d     cpc_avg_saldo      dh_val_pago_tarj_d
##  Min.   :-0.77679   Min.   :-0.47777   Min.   :-0.52372   Min.   :-0.39327  
##  1st Qu.:-0.62885   1st Qu.:-0.47777   1st Qu.:-0.43609   1st Qu.:-0.39327  
##  Median :-0.56706   Median :-0.47777   Median :-0.35621   Median :-0.39327  
##  Mean   : 0.01752   Mean   :-0.01044   Mean   : 0.05172   Mean   : 0.02682  
##  3rd Qu.: 0.26016   3rd Qu.:-0.06338   3rd Qu.:-0.01338   3rd Qu.:-0.08126  
##  Max.   : 5.44829   Max.   : 5.94527   Max.   : 8.49138   Max.   : 9.44247  
##  flag_tuvo_mora60_ult3meses flag_diaPago_1quincena flag_es_cluster_6  
##  Min.   :-0.38408           Min.   :-1.55854       Min.   :-1.008661  
##  1st Qu.:-0.38408           1st Qu.:-1.55854       1st Qu.:-1.008661  
##  Median :-0.38408           Median : 0.64158       Median : 0.991344  
##  Mean   : 0.01079           Mean   :-0.03936       Mean   : 0.002148  
##  3rd Qu.:-0.38408           3rd Qu.: 0.64158       3rd Qu.: 0.991344  
##  Max.   : 2.60348           Max.   : 0.64158       Max.   : 0.991344  
##  flag_tuvo_mora90_ult12M flax_maxDiaOtrosD_1quincena flag_encontrado_cifin
##  Min.   :-0.2625         Min.   :-1.13436            Min.   :-1.45210     
##  1st Qu.:-0.2625         1st Qu.:-1.13436            1st Qu.:-1.45210     
##  Median :-0.2625         Median : 0.88149            Median : 0.68861     
##  Mean   : 0.0003         Mean   :-0.00048            Mean   : 0.01004     
##  3rd Qu.:-0.2625         3rd Qu.: 0.88149            3rd Qu.: 0.68861     
##  Max.   : 3.8087         Max.   : 0.88149            Max.   : 0.68861

Balanceo

Se utilizará el método SMOTE para balancear las clases de la variable respuesta

temp <- train
temp$Y <- as.factor(temp$Y)
class(temp$Y)

## [1] "factor"

S <- SMOTE(Y ~ ., data = temp, perc.over = 100)
table(S$Y)

## 
##    0    1 
## 8904 8904

S$Y <- as.numeric(S$Y) - 1
table(S$Y)

## 
##    0    1 
## 8904 8904

train2 <- S

Regresión Logística

# variables comentados por VIF > 4 o PValue muy alto
modelo_logistica <- glm(
  Y ~ 
    desv_sem
  + max_mes_anterior
  + desv_trim
  + dh_cant_entradas
  + pcons_tarjeta_de_credito
  + dh_avg_dia_retiros_d
  #+ dh_cant_otros_d
  #+ dmi_max_ingreso_diario
  #+ dh_val_otros_d
  + dh_cant_pagos_d
  #+ dh_val_pagos_d
  + pc_cuotas_pagadas
  + dh_min_dia_pagos_d
  + pc_ingreso_por_rutina
  + cpc_saldo_sobre_ing
  + dh_min_dia_pago_cred_d
  + cpc_saldo_tdc
  #+ dh_cant_tras_d
  + cpc_avg_saldo
  + dh_val_pago_tarj_d
  + flag_tuvo_mora60_ult3meses
  #+ flag_diaPago_1quincena
  + flag_es_cluster_6
  + flag_tuvo_mora90_ult12M
  #+ flax_maxDiaOtrosD_1quincena
  + flag_encontrado_cifin
  , data = train2
  , family = "binomial"
)

# Prueba VIF
vif(modelo_logistica)

##                   desv_sem           max_mes_anterior 
##                   2.498924                   1.438054 
##                  desv_trim           dh_cant_entradas 
##                   2.428364                   3.044172 
##   pcons_tarjeta_de_credito       dh_avg_dia_retiros_d 
##                   1.093417                   3.082506 
##            dh_cant_pagos_d          pc_cuotas_pagadas 
##                   2.620900                   1.221223 
##         dh_min_dia_pagos_d      pc_ingreso_por_rutina 
##                   2.120878                   1.195525 
##        cpc_saldo_sobre_ing     dh_min_dia_pago_cred_d 
##                   2.036106                   2.013672 
##              cpc_saldo_tdc              cpc_avg_saldo 
##                   1.507262                   2.061050 
##         dh_val_pago_tarj_d flag_tuvo_mora60_ult3meses 
##                   1.401443                   1.691738 
##          flag_es_cluster_6    flag_tuvo_mora90_ult12M 
##                   1.202198                   1.802167 
##      flag_encontrado_cifin 
##                   1.551095

# PVaues
summary(modelo_logistica)

## 
## Call:
## glm(formula = Y ~ desv_sem + max_mes_anterior + desv_trim + dh_cant_entradas + 
##     pcons_tarjeta_de_credito + dh_avg_dia_retiros_d + dh_cant_pagos_d + 
##     pc_cuotas_pagadas + dh_min_dia_pagos_d + pc_ingreso_por_rutina + 
##     cpc_saldo_sobre_ing + dh_min_dia_pago_cred_d + cpc_saldo_tdc + 
##     cpc_avg_saldo + dh_val_pago_tarj_d + flag_tuvo_mora60_ult3meses + 
##     flag_es_cluster_6 + flag_tuvo_mora90_ult12M + flag_encontrado_cifin, 
##     family = "binomial", data = train2)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -2.0183  -1.1163   0.2137   1.0492   3.5088  
## 
## Coefficients:
##                             Estimate Std. Error z value Pr(>|z|)    
## (Intercept)                 0.054021   0.016054   3.365 0.000765 ***
## desv_sem                   -0.416694   0.032800 -12.704  < 2e-16 ***
## max_mes_anterior           -0.198661   0.026386  -7.529 5.11e-14 ***
## desv_trim                  -0.054705   0.026350  -2.076 0.037887 *  
## dh_cant_entradas            0.337139   0.028536  11.815  < 2e-16 ***
## pcons_tarjeta_de_credito   -0.108810   0.016486  -6.600 4.11e-11 ***
## dh_avg_dia_retiros_d       -0.101548   0.028240  -3.596 0.000323 ***
## dh_cant_pagos_d            -0.068394   0.025683  -2.663 0.007745 ** 
## pc_cuotas_pagadas           0.016334   0.017699   0.923 0.356092    
## dh_min_dia_pagos_d         -0.013568   0.023460  -0.578 0.563019    
## pc_ingreso_por_rutina       0.007065   0.017877   0.395 0.692695    
## cpc_saldo_sobre_ing        -0.024267   0.023083  -1.051 0.293124    
## dh_min_dia_pago_cred_d      0.075297   0.023111   3.258 0.001122 ** 
## cpc_saldo_tdc              -0.177958   0.019884  -8.950  < 2e-16 ***
## cpc_avg_saldo               0.032348   0.023794   1.360 0.173984    
## dh_val_pago_tarj_d          0.069909   0.019983   3.498 0.000468 ***
## flag_tuvo_mora60_ult3meses -0.169527   0.020893  -8.114 4.90e-16 ***
## flag_es_cluster_6           0.084413   0.017430   4.843 1.28e-06 ***
## flag_tuvo_mora90_ult12M    -0.138459   0.021116  -6.557 5.49e-11 ***
## flag_encontrado_cifin      -0.317650   0.020251 -15.685  < 2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 24687  on 17807  degrees of freedom
## Residual deviance: 22554  on 17788  degrees of freedom
## AIC: 22594
## 
## Number of Fisher Scoring iterations: 4

# Evaluar Performance
rendimiento <- function(modelo, dataset, corte)
{
  #modelo <- modelo_logistica
  #dataset <- test
  #corte <- 0.5
  R <- test$Y
  PP <- as.numeric(predict(modelo_logistica, test, type = "response"))
  # hist(PP)
  P <- ifelse(PP >= corte, 1, 0)
  tabla <- table(R,P)
  acc <- round(sum(diag(tabla)) / sum(tabla),2)
  auc <- round(as.numeric(roc(R,P)$auc),2)
  gini <- (2 * auc) - 1
  T <- as.data.frame(cbind(R,P))
  TNR <- round(nrow(subset(T, T$R == 0 & T$P == 0)) / nrow(subset(T, T$R == 0)),2)
  TPR <- round(nrow(subset(T, T$R == 1 & T$P == 1)) / nrow(subset(T, T$R == 1)),2)
  R <- as.data.frame(cbind(corte, acc, auc, gini, TNR, TPR))
  return(R)
}

# Buscar mejor punto de corte
rendimiento(modelo_logistica, test, 0.5)

## Setting levels: control = 0, case = 1

## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1   0.5 0.65 0.63 0.26 0.58 0.68

# buscar mejor punto de corte
for(i in seq(0.2, 0.8, by = 0.01))
{
  print(rendimiento(modelo_logistica, test, i))
}

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini TNR  TPR
## 1   0.2 0.7 0.54 0.08 0.1 0.97

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.21 0.7 0.54 0.08 0.11 0.97

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.22 0.7 0.54 0.08 0.11 0.97

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.23 0.7 0.54 0.08 0.12 0.97

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.24 0.7 0.55  0.1 0.13 0.96

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.25 0.7 0.55  0.1 0.13 0.96

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.26 0.7 0.55  0.1 0.14 0.95

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.27 0.7 0.55  0.1 0.15 0.95

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.28 0.7 0.56 0.12 0.17 0.94

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.29 0.7 0.56 0.12 0.18 0.94

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini TNR  TPR
## 1   0.3 0.7 0.57 0.14 0.2 0.93

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.31 0.7 0.57 0.14 0.21 0.93

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.32 0.7 0.58 0.16 0.23 0.92

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.33 0.71 0.58 0.16 0.24 0.92

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.34 0.71 0.59 0.18 0.27 0.91

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR TPR
## 1  0.35 0.71 0.6  0.2 0.29 0.9

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini TNR TPR
## 1  0.36 0.71 0.6  0.2 0.3 0.9

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1  0.37 0.71 0.6  0.2 0.32 0.88

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.38 0.7 0.61 0.22 0.34 0.87

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.39 0.7 0.61 0.22 0.35 0.86

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1   0.4 0.7 0.61 0.22 0.36 0.85

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.41 0.7 0.61 0.22 0.38 0.84

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.42 0.69 0.61 0.22 0.4 0.82

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.43 0.68 0.61 0.22 0.42 0.81

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.44 0.68 0.62 0.24 0.45 0.79

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.45 0.68 0.62 0.24 0.47 0.77

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.46 0.67 0.63 0.26 0.5 0.75

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.47 0.67 0.62 0.24 0.51 0.73

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.48 0.66 0.62 0.24 0.54 0.71

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.49 0.65 0.62 0.24 0.55 0.69

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1   0.5 0.65 0.63 0.26 0.58 0.68

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.51 0.64 0.63 0.26 0.6 0.66

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.52 0.63 0.63 0.26 0.62 0.63

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR TPR
## 1  0.53 0.62 0.63 0.26 0.65 0.6

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.54 0.61 0.63 0.26 0.68 0.58

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.55 0.6 0.62 0.24 0.69 0.56

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.56 0.59 0.62 0.24 0.71 0.53

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR TPR
## 1  0.57 0.57 0.61 0.22 0.73 0.5

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.58 0.56 0.61 0.22 0.76 0.47

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.59 0.54 0.61 0.22 0.77 0.44

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1   0.6 0.53 0.6  0.2 0.79 0.41

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.61 0.52 0.59 0.18 0.8 0.38

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.62 0.5 0.59 0.18 0.83 0.36

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.63 0.49 0.58 0.16 0.84 0.33

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.64 0.48 0.58 0.16 0.86 0.31

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.65 0.47 0.58 0.16 0.87 0.29

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.66 0.46 0.58 0.16 0.89 0.26

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.67 0.44 0.57 0.14 0.9 0.23

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.68 0.43 0.56 0.12 0.92 0.21

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.69 0.42 0.56 0.12 0.93 0.18

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1   0.7 0.4 0.55  0.1 0.94 0.16

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.71 0.39 0.54 0.08 0.95 0.13

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.72 0.38 0.53 0.06 0.96 0.11

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.73 0.36 0.53 0.06 0.97 0.08

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.74 0.35 0.52 0.04 0.97 0.07

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.75 0.35 0.52 0.04 0.98 0.06

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.76 0.33 0.51 0.02 0.99 0.03

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.77 0.33 0.51 0.02 0.99 0.03

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1  0.78 0.32 0.5    0 0.99 0.02

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1  0.79 0.32 0.5    0 0.99 0.01

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1   0.8 0.32 0.5    0 0.99 0.01

# Finalmente
resultado1 <- as.data.frame(cbind("RegLog", rendimiento(modelo_logistica, test, 0.5)))

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

names(resultado1)[1] <- "Modelo"
print(resultado1)

##   Modelo corte  acc  auc gini  TNR  TPR
## 1 RegLog   0.5 0.65 0.63 0.26 0.58 0.68

Random Forest

Se entrenó un random forest tanto con datos sin balancear como con datos balanceados a través del método SMOTE (páquete DMwR); finalmente, el 2° random forest presentó el mejor AUC con un punto de corte de 0.5

# Usar el FOR loop para buscar los mejores parámetros

# Evaluar Performance
rendimiento <- function(modelo, dataset, corte)
{
  # modelo <- RF1
  # dataset <- test
  # corte <- 0.5
  R <- test$Y
  PP <- as.numeric(predict(modelo, dataset, type = "prob")[,2])
  # hist(PP)
  P <- ifelse(PP >= corte, 1, 0)
  tabla <- table(R,P)
  acc <- round(sum(diag(tabla)) / sum(tabla),2)
  auc <- round(as.numeric(roc(R,P)$auc),2)
  gini <- (2 * auc) - 1
  T <- as.data.frame(cbind(R,P))
  TNR <- round(nrow(subset(T, T$R == 0 & T$P == 0)) / nrow(subset(T, T$R == 0)),2)
  TPR <- round(nrow(subset(T, T$R == 1 & T$P == 1)) / nrow(subset(T, T$R == 1)),2)
  R <- as.data.frame(cbind(corte, acc, auc, gini, TNR, TPR))
  return(R)
}

# random forest con data sin balancear
for(parametro in 1:1)
{
  # Modelado
  set.seed(1)
  RF1 <- randomForest(
    as.factor(Y) ~ .
    , data = train
    , ntree = 250
    , mtry = 7
    , classwt = c(1.5, 1)
  )
  
  # Evaluar performance
  print(cbind(parametro, rendimiento(RF1, test, 0.5)))
}

## Setting levels: control = 0, case = 1

## Setting direction: controls < cases

##   parametro corte  acc  auc gini  TNR  TPR
## 1         1   0.5 0.75 0.67 0.34 0.44 0.89

# random forest con data balanceada
for(parametro in 1:1)
{
  # Modelado
  set.seed(1)
  RF2 <- randomForest(
    as.factor(Y) ~ .
    , data = train2
    , ntree = 100
    , mtry = 7
    , replace = TRUE
  )
  
  # Evaluar performance
  print(cbind(parametro, rendimiento(RF2, test, 0.5)))
}

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   parametro corte  acc  auc gini TNR  TPR
## 1         1   0.5 0.71 0.71 0.42 0.7 0.71

# Evaluar todos los puntos de corte
for(i in seq(0.2, 0.8, by = 0.01))
{
  print(rendimiento(RF2, test, i))
}

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1   0.2 0.74 0.6  0.2 0.23 0.97

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.21 0.74 0.61 0.22 0.26 0.96

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.22 0.74 0.61 0.22 0.26 0.96

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.23 0.74 0.62 0.24 0.28 0.95

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.24 0.74 0.62 0.24 0.3 0.94

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.25 0.74 0.62 0.24 0.3 0.94

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.26 0.74 0.63 0.26 0.32 0.93

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.27 0.75 0.64 0.28 0.34 0.93

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.28 0.75 0.64 0.28 0.36 0.92

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.29 0.75 0.65  0.3 0.4 0.91

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR TPR
## 1   0.3 0.75 0.66 0.32 0.41 0.9

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR TPR
## 1  0.31 0.75 0.66 0.32 0.41 0.9

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.32 0.74 0.66 0.32 0.42 0.89

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.33 0.75 0.66 0.32 0.44 0.89

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.34 0.75 0.67 0.34 0.46 0.88

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.35 0.74 0.67 0.34 0.47 0.87

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.36 0.74 0.68 0.36 0.49 0.86

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.37 0.74 0.68 0.36 0.5 0.85

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.38 0.74 0.68 0.36 0.52 0.84

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.39 0.74 0.69 0.38 0.54 0.84

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1   0.4 0.74 0.69 0.38 0.56 0.82

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR TPR
## 1  0.41 0.73 0.69 0.38 0.59 0.8

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.42 0.73 0.69 0.38 0.6 0.79

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1  0.43 0.73 0.7  0.4 0.61 0.78

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1  0.44 0.73 0.7  0.4 0.61 0.78

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1  0.45 0.72 0.7  0.4 0.63 0.77

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1  0.46 0.72 0.7  0.4 0.64 0.76

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1  0.47 0.72 0.7  0.4 0.67 0.74

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1  0.48 0.71 0.7  0.4 0.68 0.72

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1  0.49 0.71 0.7  0.4 0.68 0.72

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1   0.5 0.71 0.71 0.42 0.7 0.71

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc auc gini  TNR TPR
## 1  0.51 0.7 0.7  0.4 0.71 0.7

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc auc gini  TNR  TPR
## 1  0.52 0.7 0.7  0.4 0.72 0.69

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1  0.53 0.69 0.7  0.4 0.73 0.67

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1  0.54 0.68 0.7  0.4 0.74 0.65

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1  0.55 0.68 0.7  0.4 0.76 0.64

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1  0.56 0.67 0.7  0.4 0.77 0.62

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR TPR
## 1  0.57 0.66 0.7  0.4 0.79 0.6

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.58 0.65 0.69 0.38 0.8 0.59

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.59 0.65 0.69 0.38 0.82 0.57

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1   0.6 0.64 0.69 0.38 0.82 0.55

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.61 0.63 0.68 0.36 0.83 0.53

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.62 0.63 0.68 0.36 0.83 0.53

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.63 0.62 0.68 0.36 0.85 0.52

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR TPR
## 1  0.64 0.61 0.68 0.36 0.86 0.5

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.65 0.61 0.68 0.36 0.87 0.48

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.66 0.6 0.67 0.34 0.88 0.47

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.67 0.59 0.67 0.34 0.88 0.45

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.68 0.58 0.66 0.32 0.89 0.44

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.69 0.57 0.66 0.32 0.89 0.42

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR TPR
## 1   0.7 0.56 0.65  0.3 0.9 0.4

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.71 0.55 0.64 0.28 0.9 0.38

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.72 0.54 0.64 0.28 0.91 0.37

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.73 0.53 0.63 0.26 0.91 0.35

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.74 0.52 0.62 0.24 0.92 0.33

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.75 0.51 0.62 0.24 0.92 0.32

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR TPR
## 1  0.76 0.5 0.62 0.24 0.93 0.3

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.77 0.49 0.61 0.22 0.94 0.29

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.78 0.49 0.61 0.22 0.94 0.28

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.79 0.48 0.61 0.22 0.95 0.27

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc auc gini  TNR  TPR
## 1   0.8 0.47 0.6  0.2 0.95 0.25

# finalmente:
resultado2 <- as.data.frame(cbind("RandomForest", rendimiento(RF2, test, 0.5)))

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

names(resultado2)[1] <- "Modelo"
print(resultado2)

##         Modelo corte  acc  auc gini TNR  TPR
## 1 RandomForest   0.5 0.71 0.71 0.42 0.7 0.71

# Adempas, las probabilidades obtenidas tienen la siguiente forma:
hist(as.numeric(predict(RF2, test, type = "prob")[,2]))

AdaBoost

# Preparar todo para el modelo
usar <- train2
usar$Y <- as.factor(usar$Y)
set.seed(1)

# Entrenar modelo con todas las variables por default
modeloAda <- boosting(
  Y ~ 
    desv_sem   # Importance: 33.7176260590549
  +   max_mes_anterior   # Importance: 18.8506694661822
  +   dh_cant_tras_d   # Importance: 6.66562441816714
  +   dh_min_dia_pagos_d   # Importance: 5.74571007099648
  +   flag_tuvo_mora60_ult3meses   # Importance: 5.27516596907519
  +   pc_cuotas_pagadas   # Importance: 4.94772841710666
  +   dh_cant_entradas   # Importance: 3.6451690230774
  +   flag_encontrado_cifin   # Importance: 3.51978408061676
  +   pc_ingreso_por_rutina   # Importance: 3.20238547917842
  +   desv_trim   # Importance: 3.06059257514681
  +   cpc_saldo_tdc   # Importance: 2.89533364487582
  +   dh_cant_otros_d   # Importance: 2.49716825059181
  +   dh_val_pago_tarj_d   # Importance: 1.03252217294607
  +   dh_cant_pagos_d   # Importance: 0.899084100587163
  +   cpc_avg_saldo   # Importance: 0.860227126123119
  +   dh_min_dia_pago_cred_d   # Importance: 0.790252390234587
  +   pcons_tarjeta_de_credito   # Importance: 0.623763602744435
  +   dh_val_otros_d   # Importance: 0.525760920959076
  +   dmi_max_ingreso_diario   # Importance: 0.316415754836402
  +   cpc_saldo_sobre_ing   # Importance: 0.313236673318644
  +   dh_val_pagos_d   # Importance: 0.266451837486692
  +   dh_avg_dia_retiros_d   # Importance: 0.168203660652142
  +   flag_diaPago_1quincena   # Importance: 0.104916028226669
  +   flag_es_cluster_6   # Importance: 0.0762082778154691
  +   flag_tuvo_mora90_ult12M   # Importance: 0
  +   flax_maxDiaOtrosD_1quincena   # Importance: 0
  , data = usar
)


# Imprimir importancia
modeloAda$importance # importancia de las variables

##               cpc_avg_saldo         cpc_saldo_sobre_ing 
##                  0.62945995                  0.62137479 
##               cpc_saldo_tdc                    desv_sem 
##                  2.02408780                 29.34513646 
##                   desv_trim        dh_avg_dia_retiros_d 
##                  5.06654437                  0.38408579 
##            dh_cant_entradas             dh_cant_otros_d 
##                  3.96578792                  2.53680494 
##             dh_cant_pagos_d              dh_cant_tras_d 
##                  1.01793538                 11.50905232 
##      dh_min_dia_pago_cred_d          dh_min_dia_pagos_d 
##                  1.07949399                  5.87323983 
##              dh_val_otros_d          dh_val_pago_tarj_d 
##                  0.48679549                  1.63684079 
##              dh_val_pagos_d      dmi_max_ingreso_diario 
##                  0.45164458                  0.56812405 
##      flag_diaPago_1quincena       flag_encontrado_cifin 
##                  0.04640731                  2.98015468 
##           flag_es_cluster_6  flag_tuvo_mora60_ult3meses 
##                  0.23381769                  3.30989601 
##     flag_tuvo_mora90_ult12M flax_maxDiaOtrosD_1quincena 
##                  0.23937819                  0.00000000 
##            max_mes_anterior           pc_cuotas_pagadas 
##                 18.36357557                  4.13535714 
##       pc_ingreso_por_rutina    pcons_tarjeta_de_credito 
##                  2.78858873                  0.70641620

# Evaluar Performance del modelo
rendimiento <- function(modelo, dataset, corte)
{
  # modelo <- modeloAda
  # dataset <- test
  # corte <- 0.5
  R <- test$Y
  PP <- (as.numeric(predict(modelo, dataset)$prob[,2]))
  # hist(PP)
  P <- ifelse(PP >= corte, 1, 0)
  tabla <- table(R,P)
  acc <- round(sum(diag(tabla)) / sum(tabla),2)
  auc <- round(as.numeric(roc(R,P)$auc),2)
  gini <- (2 * auc) - 1
  T <- as.data.frame(cbind(R,P))
  TNR <- round(nrow(subset(T, T$R == 0 & T$P == 0)) / nrow(subset(T, T$R == 0)),2)
  TPR <- round(nrow(subset(T, T$R == 1 & T$P == 1)) / nrow(subset(T, T$R == 1)),2)
  R <- as.data.frame(cbind(corte, acc, auc, gini, TNR, TPR))
  return(R)
}


# Modelo omitiendo algunas variables
modeloAda <- boosting(
  Y ~ 
    desv_sem   # Importance: 33.7176260590549
  +   max_mes_anterior   # Importance: 18.8506694661822
  +   dh_cant_tras_d   # Importance: 6.66562441816714
  +   dh_min_dia_pagos_d   # Importance: 5.74571007099648
  +   flag_tuvo_mora60_ult3meses   # Importance: 5.27516596907519
  +   pc_cuotas_pagadas   # Importance: 4.94772841710666
  +   dh_cant_entradas   # Importance: 3.6451690230774
  +   flag_encontrado_cifin   # Importance: 3.51978408061676
  +   pc_ingreso_por_rutina   # Importance: 3.20238547917842
  +   desv_trim   # Importance: 3.06059257514681
  +   cpc_saldo_tdc   # Importance: 2.89533364487582
  +   dh_cant_otros_d   # Importance: 2.49716825059181
  +   dh_val_pago_tarj_d   # Importance: 1.03252217294607
  #+   dh_cant_pagos_d   # Importance: 0.899084100587163
  #+   cpc_avg_saldo   # Importance: 0.860227126123119
  #+   dh_min_dia_pago_cred_d   # Importance: 0.790252390234587
  #+   pcons_tarjeta_de_credito   # Importance: 0.623763602744435
  #+   dh_val_otros_d   # Importance: 0.525760920959076
  #+   dmi_max_ingreso_diario   # Importance: 0.316415754836402
  #+   cpc_saldo_sobre_ing   # Importance: 0.313236673318644
  #+   dh_val_pagos_d   # Importance: 0.266451837486692
  #+   dh_avg_dia_retiros_d   # Importance: 0.168203660652142
  #+   flag_diaPago_1quincena   # Importance: 0.104916028226669
  #+   flag_es_cluster_6   # Importance: 0.0762082778154691
  #+   flag_tuvo_mora90_ult12M   # Importance: 0
  #+   flax_maxDiaOtrosD_1quincena   # Importance: 0
  , data = usar
  , mfinal = 10
  , boos = TRUE
)

# Performance del modelo
rendimiento(modeloAda, test, 0.5)

## Setting levels: control = 0, case = 1

## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1   0.5 0.66 0.64 0.28 0.59 0.69

# Buscar mejor punto de corte
for(i in seq(0.2, 0.8, by = 0.01))
{
  print(rendimiento(modeloAda, test, i))
}

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1   0.2 0.71 0.58 0.16 0.22 0.93

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.21 0.71 0.58 0.16 0.25 0.92

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.22 0.71 0.58 0.16 0.25 0.92

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.23 0.7 0.59 0.18 0.26 0.91

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR TPR
## 1  0.24 0.7 0.59 0.18 0.27 0.9

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini TNR  TPR
## 1  0.25 0.7 0.59 0.18 0.3 0.89

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc auc gini  TNR  TPR
## 1  0.26 0.7 0.6  0.2 0.31 0.88

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc auc gini  TNR  TPR
## 1  0.27 0.7 0.6  0.2 0.33 0.87

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc auc gini  TNR  TPR
## 1  0.28 0.7 0.6  0.2 0.34 0.87

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc auc gini  TNR  TPR
## 1  0.29 0.7 0.6  0.2 0.35 0.86

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1   0.3 0.7 0.61 0.22 0.36 0.85

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.31 0.7 0.61 0.22 0.37 0.85

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.32 0.7 0.61 0.22 0.38 0.84

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.33 0.69 0.61 0.22 0.39 0.84

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.34 0.69 0.62 0.24 0.41 0.83

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.35 0.69 0.62 0.24 0.41 0.82

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.36 0.69 0.62 0.24 0.43 0.81

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR TPR
## 1  0.37 0.69 0.62 0.24 0.45 0.8

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR TPR
## 1  0.38 0.69 0.62 0.24 0.45 0.8

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.39 0.69 0.63 0.26 0.46 0.79

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1   0.4 0.68 0.63 0.26 0.48 0.77

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.41 0.67 0.63 0.26 0.5 0.76

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.42 0.68 0.63 0.26 0.51 0.75

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.43 0.68 0.63 0.26 0.52 0.75

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.44 0.67 0.63 0.26 0.53 0.73

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.45 0.67 0.63 0.26 0.54 0.73

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.46 0.67 0.63 0.26 0.54 0.73

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.47 0.67 0.64 0.28 0.56 0.72

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.48 0.66 0.64 0.28 0.56 0.71

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR TPR
## 1  0.49 0.66 0.64 0.28 0.57 0.7

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1   0.5 0.66 0.64 0.28 0.59 0.69

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.51 0.65 0.64 0.28 0.6 0.68

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.52 0.65 0.64 0.28 0.6 0.67

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.53 0.65 0.64 0.28 0.61 0.67

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.54 0.64 0.64 0.28 0.62 0.66

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.55 0.64 0.64 0.28 0.63 0.64

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.56 0.63 0.64 0.28 0.64 0.63

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.57 0.63 0.64 0.28 0.66 0.61

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR TPR
## 1  0.58 0.62 0.64 0.28 0.68 0.6

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.59 0.62 0.64 0.28 0.69 0.59

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1   0.6 0.62 0.64 0.28 0.69 0.59

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.61 0.61 0.63 0.26 0.7 0.57

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1  0.62 0.61 0.63 0.26 0.7 0.56

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.63 0.6 0.64 0.28 0.72 0.55

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.64 0.59 0.63 0.26 0.74 0.52

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.65 0.59 0.63 0.26 0.74 0.52

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.66 0.58 0.63 0.26 0.75 0.51

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.67 0.57 0.62 0.24 0.75 0.49

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.68 0.56 0.62 0.24 0.77 0.47

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.69 0.56 0.62 0.24 0.79 0.45

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1   0.7 0.55 0.62 0.24 0.8 0.44

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.71 0.55 0.62 0.24 0.81 0.42

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.72 0.54 0.62 0.24 0.81 0.42

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR TPR
## 1  0.73 0.54 0.61 0.22 0.82 0.4

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte acc  auc gini  TNR  TPR
## 1  0.74 0.5 0.59 0.18 0.86 0.33

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.75 0.49 0.59 0.18 0.86 0.32

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.76 0.49 0.59 0.18 0.86 0.32

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.77 0.48 0.59 0.18 0.87 0.31

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.78 0.48 0.58 0.16 0.87 0.29

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini  TNR  TPR
## 1  0.79 0.47 0.58 0.16 0.87 0.29

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

##   corte  acc  auc gini TNR  TPR
## 1   0.8 0.46 0.58 0.16 0.9 0.26

resultado3 <- cbind("AdaBoost", rendimiento(modeloAda, test, 0.56))

## Setting levels: control = 0, case = 1
## Setting direction: controls < cases

names(resultado3)[1] <- "Modelo"
resultado3

##     Modelo corte  acc  auc gini  TNR  TPR
## 1 AdaBoost  0.56 0.63 0.64 0.28 0.64 0.63

Torneo de modelos

El modelo ganador es un Random Forest usando la data balanceada, ntree = 100 y mtry = 7. Este modelo nos da un AUC de 0.71 y un GINI de 0.42. Dado que el AUC es >= 0.70 el modelo se considera aceptable y viable para su utilización, pero se recomiendo buscar un AUC >= 0.80 a través de:

• Agregar nuevas variables
• Probar otras transformaciones de variables
• Usar algoritmos más complejos como XGboost o CatBoost
• Finalmente, se puede probar un análisis factorial para el diseño de experimentos

# Finalmente...

rbind(
  resultado1,
  resultado2,
  resultado3
)

##         Modelo corte  acc  auc gini  TNR  TPR
## 1       RegLog  0.50 0.65 0.63 0.26 0.58 0.68
## 2 RandomForest  0.50 0.71 0.71 0.42 0.70 0.71
## 3     AdaBoost  0.56 0.63 0.64 0.28 0.64 0.63

OOT dataset y creación de base_prueba evaluado

original <- read.csv("Base_prueba.csv", header = TRUE)
dim(original) # cargadas 20,000 filas y 124 columnas

## [1] 1000  123

data <- original # sobre este dataset se hará el feature engineer

# Omitir datos atípicos

A <- nrow(data)
data <- subset(data, data$pc_transaccional <= 136903394) 
data <- subset(data, data$dmi_max_egreso_diario <= 85362328)
data <- subset(data, data$dmi_max_ingreso_diario <= 104909227.3)
data <- subset(data, data$dh_val_otros_d <= 9998335.80)
data <- subset(data, data$dmi_ingreso_total_mes <= 205163705.0)
data <- subset(data, data$dh_val_pagos_d <= 50357360.00)
data <- subset(data, data$pc_gasto_familiar <= 26582180.19)
data <- subset(data, data$pc_cuotas_pagadas <= 8156850.0)
data <- subset(data, data$dmi_egreso_total_mes <= 209829245.0)
data <- subset(data, data$dh_val_salidas <= 217011100)
data <- subset(data, data$cpc_sum_saldo <= 220905770.00)
data <- subset(data, data$cpc_saldo_sobre_ing <= 20.307294513)
data <- subset(data, data$cpc_saldo_tdc <= 43186883.96)
data <- subset(data, data$dh_cant_tras_d <= 31)
data <- subset(data, data$cpc_avg_saldo <= 106884393.50)
data <- subset(data, data$dh_val_pago_tarj_d <= 10560315.0)
data <- subset(data, data$pc_transaccional <= 92269310.0)
B <- nrow(data)

paste(A-B,"registros omitidos",sep = " ")

## [1] "98 registros omitidos"

# Exploración y Transformación de variables

names(data)[1] # max_trim: máximo días de mora trimestre anterior

## [1] "max_trim"

hist(data$max_trim)

hist(sqrt(data$max_trim)) # variable normalizada con sqrt

max_trim <- sqrt(data$max_trim)

names(data)[2] # max_sem: máximo días de mora en el semestre anterior

## [1] "max_sem"

hist(data$max_sem)

hist(sqrt(data$max_sem))

max_sem <- data$max_sem

names(data)[3] # desv_sem: desviación estándar del máximo de moras en el semestre anterior

## [1] "desv_sem"

hist(data$desv_sem)

hist(sqrt(data$desv_sem))

desv_sem <- data$desv_sem

names(data)[4] # prom_bim: promedio del máximo de moras en el bimestre anterior

## [1] "prom_bim"

hist(data$prom_bim)

hist(sqrt(data$prom_bim))

prom_bim <- sqrt(data$prom_bim)

names(data)[5] # max_mes_anterior: días de mora máximo en el mes anterior

## [1] "max_mes_anterior"

hist(data$max_mes_anterior)

hist(sqrt(data$max_mes_anterior))

max_mes_anterior <- data$max_mes_anterior

names(data)[6] # prom_mes_anterior: Promedio de los dias de mora en el mes anterior

## [1] "prom_mes_anterior"

hist(data$prom_mes_anterior)

hist(sqrt(data$prom_mes_anterior))

prom_mes_anterior <- sqrt(data$prom_mes_anterior)

names(data)[7] # prom_sem: promedio del máximo de moras en el semestre anterior

## [1] "prom_sem"

hist(data$prom_sem)

hist(sqrt(data$prom_sem))

prom_sem <- data$prom_sem

names(data)[8] # max_bim: máximo días de mora en el bimestre anterior

## [1] "max_bim"

hist(data$max_bim)

hist(sqrt(data$max_bim))

max_bim <- data$max_bim

names(data)[9] # mejor_gestion: Mejor gestion realizada

## [1] "mejor_gestion"

table(data$mejor_gestion)

## 
##   0  15 
## 901   1

names(data)[10] # prom_trim: promedio del máximo de moras en el trimestre anterior

## [1] "prom_trim"

hist(data$prom_trim)

hist(sqrt(data$prom_trim))

prom_trim <- sqrt(data$prom_trim)

names(data)[11] # pc_cant_moras_30_ult_12_meses: Cantidad de moras 30 en los últimos 12 meses <= 5 o vacío.

## [1] "pc_cant_moras_30_ult_12_meses"

summary(data$pc_cant_moras_30_ult_12_meses)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##  0.0000  0.0000  0.0000  0.8758  1.0000 12.0000

table(data$pc_cant_moras_30_ult_12_meses)

## 
##   0   1   2   3   4   6   7   8   9  10  11  12 
## 639 124  24  31  32  18  12   3  12   5   1   1

pc_cant_moras_30_ult_12_meses <- data$pc_cant_moras_30_ult_12_meses

names(data)[12] # desv_trim: desviación estándar del máximo de moras en el trimestre  anterior

## [1] "desv_trim"

hist(data$desv_trim)

hist(sqrt(data$desv_trim))

desv_trim <- sqrt(data$desv_trim)

names(data)[13] # nro_gestiones: Numero de gestiones realizadas

## [1] "nro_gestiones"

table(data$nro_gestiones)

## 
##   0   2 
## 901   1

names(data)[14] # desv_bim: desviación estándar del máximo días de mora en el bimestre anterior

## [1] "desv_bim"

hist(data$desv_bim)

hist(sqrt(data$desv_bim))

desv_bim <- sqrt(data$desv_bim)

names(data)[15] # pc_cant_moras_30_ult_3_meses: Cantidad de moras 30 en los últimos 3 meses = 0 o vacío.

## [1] "pc_cant_moras_30_ult_3_meses"

summary(data$pc_cant_moras_30_ult_3_meses)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##  0.0000  0.0000  0.0000  0.3869  0.0000  3.0000

hist(data$pc_cant_moras_30_ult_3_meses)

names(data)[16] # dh_cant_entradas: cantidad de trasacciones de ingreso de dinero tuvo en el mes anterior

## [1] "dh_cant_entradas"

summary(data$dh_cant_entradas)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    0.00    0.00    6.00   11.59   15.00  129.00

hist(data$dh_cant_entradas)

hist(sqrt(data$dh_cant_entradas))

dh_cant_entradas <- sqrt(data$dh_cant_entradas)

names(data)[17] # pc_tiem_1er_prod_abierto_total: Tiempo total de producto abierto

## [1] "pc_tiem_1er_prod_abierto_total"

table(data$pc_tiem_1er_prod_abierto_total)

## 
##   0  34  63  65  66 122 
## 897   1   1   1   1   1

hist(data$pc_tiem_1er_prod_abierto_total)

names(data)[18] # pc_cant_moras_60_ult_12_meses: Cantidad de moras 60 en los últimos 12 meses <= 1 o vacío.

## [1] "pc_cant_moras_60_ult_12_meses"

table(data$pc_cant_moras_60_ult_12_meses)

## 
##   0   1   2   3   4   5   6   7   8   9  12 
## 778  27  35  23   5   5   6  11   6   5   1

hist(data$pc_cant_moras_60_ult_12_meses) # se optará por hacerla binaria

x <- ifelse(data$pc_cant_moras_60_ult_12_meses >= 1, 1, 0)
table(x)

## x
##   0   1 
## 778 124

flag_mora60_ult12meses <- x

names(data)[19] # gestiones_eficaces: Cantidad de gestiones eficaces

## [1] "gestiones_eficaces"

table(data$gestiones_eficaces)

## 
##   0   1 
## 901   1

names(data)[20] # pc_transaccional: Ingreso de acuerdo al estimador transaccional del cliente

## [1] "pc_transaccional"

summary(data$pc_transaccional)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0  3520425  8811787 15070019 20470286 87552251

boxplot(data$pc_transaccional) # valores atípicos en el lado superior de la variable

quantile(data$pc_transaccional, seq(0, 1, by = 0.01))

##       0%       1%       2%       3%       4%       5%       6%       7% 
##        0        0        0        0        0        0        0        0 
##       8%       9%      10%      11%      12%      13%      14%      15% 
##        0        0        0        0        0    12000   125570   434500 
##      16%      17%      18%      19%      20%      21%      22%      23% 
##   931700  1232876  1583097  2327651  2390242  2598273  2905406  3127205 
##      24%      25%      26%      27%      28%      29%      30%      31% 
##  3327102  3520425  3654249  3880363  3900000  4154062  4341266  4785855 
##      32%      33%      34%      35%      36%      37%      38%      39% 
##  4785855  5056821  5270000  5446336  5684445  5822250  6001779  6187601 
##      40%      41%      42%      43%      44%      45%      46%      47% 
##  6392479  6392479  6487250  6531800  6707020  6757348  6964516  7264240 
##      48%      49%      50%      51%      52%      53%      54%      55% 
##  7654414  8280285  8811787  9280070  9411475  9826828 10069905 10765247 
##      56%      57%      58%      59%      60%      61%      62%      63% 
## 11115000 11212284 11863116 12310765 12733000 12796336 13160097 13700000 
##      64%      65%      66%      67%      68%      69%      70%      71% 
## 14952138 15247194 15386678 15532112 16000000 16566361 16929221 18004000 
##      72%      73%      74%      75%      76%      77%      78%      79% 
## 18543000 18662584 20000000 20470286 20470286 21185026 21500000 22029301 
##      80%      81%      82%      83%      84%      85%      86%      87% 
## 24256619 24256619 24613534 25494191 27920966 28156019 29246260 30636482 
##      88%      89%      90%      91%      92%      93%      94%      95% 
## 33093800 37612700 37940660 42301382 42852690 45546580 54933792 55731375 
##      96%      97%      98%      99%     100% 
## 59325287 64094730 78140440 78196633 87552251

x <- subset(data$pc_transaccional, data$pc_transaccional <= 92269310.0)
hist(x)

pc_transaccional <- data$pc_transaccional

names(data)[21] # dh_max_dia_entradas: Ultimo dia en que recibio alguna transaccion de ingreso de dinero

## [1] "dh_max_dia_entradas"

table(data$dh_max_dia_entradas)

## 
##   0   1   2   3   4   5   6   7   9  10  11  12  13  14  15  16  18  20  21  22 
## 291   2   3   1   2   1   2   1   1   1   1   1   4   6   2   1   5   2   2   5 
##  23  24  25  26  27  28  29  30  31 
##  11   8  19   8  23  78  24 145 252

hist(data$dh_max_dia_entradas) # se volverá binaria

x <- ifelse(data$dh_max_dia_entradas <= 15, 1, 0)
table(x)

## x
##   0   1 
## 583 319

flag_ultima_entrada_1quincena <- x

names(data)[22] # pc_cupo_entidad: Cupo de las tarjetas de crédito en el banco

## [1] "pc_cupo_entidad"

summary(data$pc_cupo_entidad)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0        0    21508        0 17000000

hist(data$pc_cupo_entidad)

table(ifelse(data$pc_cupo_entidad >= 1, 1, 0))

## 
##   0   1 
## 899   3

names(data)[23] # pc_cuotas_como_ppal: Cuotas pagadas como principal

## [1] "pc_cuotas_como_ppal"

summary(data$pc_cuotas_como_ppal)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0   18017       0 6103800

hist(data$pc_cuotas_como_ppal)

names(data)[24] # dh_val_entradas: Valor total de los ingresos tomados en el mes anterior

## [1] "dh_val_entradas"

summary(data$dh_val_entradas)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0   5685000  17657538  25060713 181987436

hist(data$dh_val_entradas)

hist(sqrt(data$dh_val_entradas))

dh_val_entradas <- sqrt(data$dh_val_entradas)

names(data)[25] # pc_cant_moras_90_ult_12_meses: Cantidad de moras 90 o superiores en los últimos 12 meses = 0 o vacío.

## [1] "pc_cant_moras_90_ult_12_meses"

table(data$pc_cant_moras_90_ult_12_meses)

## 
##   0   1   2   3   4   5   6   7  12 
## 814  27  22   7   4  13   3  11   1

hist(data$pc_cant_moras_90_ult_12_meses)

hist(sqrt(data$pc_cant_moras_90_ult_12_meses))

names(data)[26] # dh_max_dia_salidas

## [1] "dh_max_dia_salidas"

# prefiero quedarme con la misma variable pero para el mes actual

names(data)[27] # pc_cant_moras_60_ult_3_meses: Cantidad de moras 60 en los últimos 3 meses = 0 o vacío.

## [1] "pc_cant_moras_60_ult_3_meses"

table(data$pc_cant_moras_60_ult_3_meses)

## 
##   0   1   2   3 
## 800  33  24  45

table(ifelse(data$pc_cant_moras_60_ult_3_meses >= 1, 1, 0))

## 
##   0   1 
## 800 102

flag_tuvo_mora60_ult3meses <- ifelse(data$pc_cant_moras_60_ult_3_meses >= 1, 1, 0)

names(data)[28] # pc_cuota_tarjeta_de_credito: Cuota de tarjeta de crédito  reportada por CIFIN

## [1] "pc_cuota_tarjeta_de_credito"

summary(data$pc_cuota_tarjeta_de_credito)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0    9882       0 4726800

hist(data$pc_cuota_tarjeta_de_credito)

table(ifelse(data$pc_cuota_tarjeta_de_credito >= 1, 1, 0))

## 
##   0   1 
## 897   5

names(data)[29] # cp_inicial_menos_saldo: valor inicial menos el saldo en el mes anterior por producto

## [1] "cp_inicial_menos_saldo"

summary(data$cp_inicial_menos_saldo)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[30] # pc_peor_estado_act_cta_aho: Peor estado cuenta ahorro

## [1] "pc_peor_estado_act_cta_aho"

summary(data$pc_peor_estado_act_cta_aho)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##  0.0000  0.0000  0.0000  0.0133  0.0000  3.0000

hist(data$pc_peor_estado_act_cta_aho)

names(data)[31] # dia_pago: Dia de pago de la obligacion en el mes

## [1] "dia_pago"

table(data$dia_pago)

## 
##   1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  17  18  19  20 
##  38  12 198  11  13   8  12   9  14  13  19   9  15  11 246  17 103  26   9  10 
##  21  22  23  24  25  26  27  28  29  30  31 
##   5  10  10  14   8  14   9   4  17  17   1

x <- ifelse(data$dia_pago <= 15, 1, 0)
table(x)

## x
##   0   1 
## 274 628

flag_diaPago_1quincena <- x

names(data)[32] # cp_cuotas_falta: Cantidad de cuotas faltantes

## [1] "cp_cuotas_falta"

table(data$cp_cuotas_falta)

## 
##   0 
## 902

names(data)[33] # pcons_tarjeta_de_credito: Es el producto una tarjeta de credito

## [1] "pcons_tarjeta_de_credito"

table(data$pcons_tarjeta_de_credito)

## 
##   0   1 
## 607 295

pcons_tarjeta_de_credito <- data$pcons_tarjeta_de_credito

names(data)[34] # pc_cifin: Ingreso del cliente de acuerdo a CIFIN

## [1] "pc_cifin"

summary(data$pc_cifin)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0   21787       0 7505630

hist(data$pc_cifin)

names(data)[35] # gestiones_prod: Numero de gestiones productivas

## [1] "gestiones_prod"

table(data$gestiones_prod)

## 
##   0   1 
## 901   1

names(data)[36] # pcons_vehiculos_sufi: Es el producto un vehiculo sufi

## [1] "pcons_vehiculos_sufi"

table(data$pcons_vehiculos_sufi)

## 
##   0   1 
## 836  66

names(data)[37] # cluster_recod: segmento cluster

## [1] "cluster_recod"

table(data$cluster_recod)

## 
##   1   6   7  13  16  19 
##   8 458 246  82  70  38

x <- ifelse(data$cluster_recod == 6, 1, 0)
table(x)

## x
##   0   1 
## 444 458

flag_es_cluster_6 <- x

names(data)[38] # dh_avg_dia_retiros_d: Dia promedio del mes en el que realiza los retiros 

## [1] "dh_avg_dia_retiros_d"

hist(data$dh_avg_dia_retiros_d)

dh_avg_dia_retiros_d <- data$dh_avg_dia_retiros_d

names(data)[39] # dmi_max_egreso_diario: Maximo egreso en un dia del mes anterior 

## [1] "dmi_max_egreso_diario"

summary(data$dmi_max_egreso_diario)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0  2509288  5687549  8127384 82421516

hist(data$dmi_max_egreso_diario)

quantile(data$dmi_max_egreso_diario, seq(0,1,by = 0.01))

##          0%          1%          2%          3%          4%          5% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##          6%          7%          8%          9%         10%         11% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         12%         13%         14%         15%         16%         17% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         18%         19%         20%         21%         22%         23% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         24%         25%         26%         27%         28%         29% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         30%         31%         32%         33%         34%         35% 
##        0.00        0.00        0.00        0.00        0.00    11399.35 
##         36%         37%         38%         39%         40%         41% 
##    57635.56    76406.00   282165.00   546344.40   848246.40  1120309.30 
##         42%         43%         44%         45%         46%         47% 
##  1278535.00  1506000.00  1586002.00  1800000.00  1808016.60  2005671.00 
##         48%         49%         50%         51%         52%         53% 
##  2071476.36  2412097.00  2509287.50  2788249.00  2999088.00  3012780.00 
##         54%         55%         56%         57%         58%         59% 
##  3329386.00  3514000.00  3614412.00  3790122.86  3917983.00  3993058.06 
##         60%         61%         62%         63%         64%         65% 
##  4314267.60  4329499.00  4760574.00  5179535.80  5434157.40  5609550.35 
##         66%         67%         68%         69%         70%         71% 
##  6002005.00  6191910.48  6386272.00  6571371.65  6966758.20  7028000.00 
##         72%         73%         74%         75%         76%         77% 
##  7321076.92  7619654.00  7970963.00  8127383.75  8774232.00  8875431.00 
##         78%         79%         80%         81%         82%         83% 
##  9546222.00  9937420.51 10054365.60 10794511.00 11621264.60 11949578.85 
##         84%         85%         86%         87%         88%         89% 
## 12373146.00 13004433.00 13213167.00 13285191.00 14333466.00 15038382.25 
##         90%         91%         92%         93%         94%         95% 
## 15383235.00 15568797.00 16518175.44 18115766.45 20069227.10 20599199.10 
##         96%         97%         98%         99%        100% 
## 22645335.40 24225138.55 31233119.30 35903859.00 82421516.00

x <- subset(data$dmi_max_egreso_diario, data$dmi_max_egreso_diario <= 85362328)
hist(x)

dmi_max_egreso_diario <- data$dmi_max_egreso_diario

names(data)[40] # cpc_max_proc_deuda: Máximo del porcentaje de la deuda en el mes anterior

## [1] "cpc_max_proc_deuda"

summary(data$cpc_max_proc_deuda)

##       Min.    1st Qu.     Median       Mean    3rd Qu.       Max. 
## -28.182000   0.000000   0.003515   0.158809   0.847877   1.272912

names(data)[41] # dh_cant_otros_d: Cantidad de trasacciones de salida por concepto de otros

## [1] "dh_cant_otros_d"

summary(data$dh_cant_otros_d)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   0.000   0.000   1.000   3.204   5.000  28.000

hist(data$dh_cant_otros_d)

hist(sqrt(data$dh_cant_otros_d))

dh_cant_otros_d <- sqrt(data$dh_cant_otros_d)

names(data)[42] # pc_cont_30_lt_12m_tot_sf: Cantidad mora 30 últimos 12 meses sector financiero

## [1] "pc_cont_30_lt_12m_tot_sf"

summary(data$pc_cont_30_lt_12m_tot_sf)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

hist(data$pc_cont_30_lt_12m_tot_sf)

table(data$pc_cont_30_lt_12m_tot_sf)

## 
##   0 
## 902

names(data)[43] # pc_cant_mora90_ult_12m_total

## [1] "pc_cant_mora90_ult_12m_total"

table(data$pc_cant_mora90_ult_12m_total)

## 
##   0   1   2   3   4   5   6   7  12 
## 856  16   2   1   4   8   3  11   1

table(ifelse(data$pc_cant_mora90_ult_12m_total >= 1, 1, 0))

## 
##   0   1 
## 856  46

x <- ifelse(data$pc_cant_mora90_ult_12m_total >= 1, 1, 0)
flag_tuvo_mora90_ult12M <- x

names(data)[44] # dmi_max_ingreso_diario: Maximo ingreso en un dia del mes anterior 

## [1] "dmi_max_ingreso_diario"

summary(data$dmi_max_ingreso_diario)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0  2534003  6832447 10000000 99410001

hist(data$dmi_max_ingreso_diario)

quantile(data$dmi_max_ingreso_diario, seq(0, 1, by = 0.01))

##          0%          1%          2%          3%          4%          5% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##          6%          7%          8%          9%         10%         11% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         12%         13%         14%         15%         16%         17% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         18%         19%         20%         21%         22%         23% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         24%         25%         26%         27%         28%         29% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         30%         31%         32%         33%         34%         35% 
##        0.00        0.00        0.00        0.00        0.00        2.35 
##         36%         37%         38%         39%         40%         41% 
##      309.00      615.78    50000.00   354796.22   835472.40  1000000.00 
##         42%         43%         44%         45%         46%         47% 
##  1200000.00  1485150.36  1648761.12  1944331.70  2045478.26  2123515.51 
##         48%         49%         50%         51%         52%         53% 
##  2337440.00  2402294.67  2534003.00  2984507.57  3000006.56  3215770.00 
##         54%         55%         56%         57%         58%         59% 
##  3460925.00  3622502.05  3843849.28  3900211.00  4207500.00  4407615.09 
##         60%         61%         62%         63%         64%         65% 
##  4651730.20  5067101.17  5500000.00  5728921.55  6019212.60  6209045.40 
##         66%         67%         68%         69%         70%         71% 
##  6399660.00  6875000.00  7000000.00  7200000.00  7348904.00  8000000.00 
##         72%         73%         74%         75%         76%         77% 
##  8378862.00  9052866.00  9986436.52 10000000.00 10000000.00 10000000.00 
##         78%         79%         80%         81%         82%         83% 
## 10257019.00 10638016.52 11216000.00 11974437.86 13164000.00 13445811.23 
##         84%         85%         86%         87%         88%         89% 
## 13700001.00 14357715.70 14950272.84 15830394.00 17779183.08 18490095.00 
##         90%         91%         92%         93%         94%         95% 
## 19175443.00 19495802.00 20801484.44 23791026.00 25000000.00 26902125.00 
##         96%         97%         98%         99%        100% 
## 33287967.36 35678300.00 39273879.36 41724001.00 99410001.00

x <- subset(data$dmi_max_ingreso_diario, data$dmi_max_ingreso_diario <= 104909227.3)
hist(x)

dmi_max_ingreso_diario <- data$dmi_max_ingreso_diario

names(data)[45] # dh_val_otros_d: Valor total de las trasacciones de salida por concepto de otros en un mes

## [1] "dh_val_otros_d"

summary(data$dh_val_otros_d)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0   12586  802098  697294 9777278

hist(data$dh_val_otros_d)

quantile(data$dh_val_otros_d, seq(0,1,by=0.01))

##         0%         1%         2%         3%         4%         5%         6% 
##       0.00       0.00       0.00       0.00       0.00       0.00       0.00 
##         7%         8%         9%        10%        11%        12%        13% 
##       0.00       0.00       0.00       0.00       0.00       0.00       0.00 
##        14%        15%        16%        17%        18%        19%        20% 
##       0.00       0.00       0.00       0.00       0.00       0.00       0.00 
##        21%        22%        23%        24%        25%        26%        27% 
##       0.00       0.00       0.00       0.00       0.00       0.00       0.00 
##        28%        29%        30%        31%        32%        33%        34% 
##       0.00       0.00       0.00       0.00       0.00       0.00       0.00 
##        35%        36%        37%        38%        39%        40%        41% 
##       0.00       0.00       0.00       0.00       0.00       0.00       0.00 
##        42%        43%        44%        45%        46%        47%        48% 
##       0.00       0.00       0.00       0.00       0.00    1610.00    2756.40 
##        49%        50%        51%        52%        53%        54%        55% 
##   10100.00   12586.00   13271.00   19601.04   20200.00   28692.00   38891.20 
##        56%        57%        58%        59%        60%        61%        62% 
##   49211.00   56686.15   60420.00   68691.68   76824.00  119253.19  133905.00 
##        63%        64%        65%        66%        67%        68%        69% 
##  174475.03  201200.00  233389.00  271140.00  309311.93  332894.00  349316.00 
##        70%        71%        72%        73%        74%        75%        76% 
##  407576.50  452845.97  483246.00  513799.14  596183.08  697293.75  713901.68 
##        77%        78%        79%        80%        81%        82%        83% 
##  827700.00  928996.18 1033514.61 1102974.80 1127520.00 1271960.24 1384444.00 
##        84%        85%        86%        87%        88%        89%        90% 
## 1648990.52 1716832.70 1872117.00 2066960.98 2388746.72 2564671.00 2652146.00 
##        91%        92%        93%        94%        95%        96%        97% 
## 2956094.31 3320259.56 3663815.21 4115774.00 5073191.50 5825378.92 6362258.37 
##        98%        99%       100% 
## 7211642.08 7645053.36 9777278.00

x <- subset(data$dh_val_otros_d, data$dh_val_otros_d <= 9998335.80)
hist(x)

dh_val_otros_d <- data$dh_val_otros_d

names(data)[46] # pc_ingreso_final: Ingreso de final del cliente

## [1] "pc_ingreso_final"

summary(data$pc_ingreso_final)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0  3315000  8362472 14053377 18708862 87552251

hist(data$pc_ingreso_final)

pc_ingreso_final <- data$pc_ingreso_final

names(data)[47] # dh_cant_pagos_d: Cantidad de pagos de salidas tuvo en el mes anterior

## [1] "dh_cant_pagos_d"

summary(data$dh_cant_pagos_d)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   0.000   0.000   3.500   5.346   8.000  35.000

hist(data$dh_cant_pagos_d)

dh_cant_pagos_d <- data$dh_cant_pagos_d

names(data)[48] # dmi_ingreso_total_mes: Ingreso total del mes anterior

## [1] "dmi_ingreso_total_mes"

summary(data$dmi_ingreso_total_mes)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0   5685000  17657538  25060713 181987436

hist(data$dmi_ingreso_total_mes)

quantile(data$dmi_ingreso_total_mes, seq(0, 1, by = 0.01))

##           0%           1%           2%           3%           4%           5% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##           6%           7%           8%           9%          10%          11% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          12%          13%          14%          15%          16%          17% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          18%          19%          20%          21%          22%          23% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          24%          25%          26%          27%          28%          29% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          30%          31%          32%          33%          34%          35% 
##         0.00         0.00         0.00         0.00         0.00         3.70 
##          36%          37%          38%          39%          40%          41% 
##       317.28      1592.74     53032.00    520933.81   1001004.40   1724907.50 
##          42%          43%          44%          45%          46%          47% 
##   2240001.00   2520715.00   3185149.60   3571411.35   4411157.02   4817541.00 
##          48%          49%          50%          51%          52%          53% 
##   5160000.00   5461628.03   5685000.00   6531800.00   7248359.04   7793376.50 
##          54%          55%          56%          57%          58%          59% 
##   8270000.00   8530406.00   9188379.16   9658108.00  10416391.60  11189637.61 
##          60%          61%          62%          63%          64%          65% 
##  11948978.60  12418588.01  13000000.00  13673026.68  14513000.00  15067301.05 
##          66%          67%          68%          69%          70%          71% 
##  16128330.00  16748600.00  18163873.72  18547426.92  20132597.40  21500000.00 
##          72%          73%          74%          75%          76%          77% 
##  22000000.00  22528017.00  23377940.20  25060713.00  27384078.44  28400354.79 
##          78%          79%          80%          81%          82%          83% 
##  30557407.96  31241943.45  31844647.40  31944284.00  34174945.00  36977650.33 
##          84%          85%          86%          87%          88%          89% 
##  39152130.00  43148753.00  46185407.00  48841121.03  52888577.60  53816089.00 
##          90%          91%          92%          93%          94%          95% 
##  56323174.00  60959669.06  66736925.00  67675419.00  68100000.00  71880165.00 
##          96%          97%          98%          99%         100% 
##  73308273.00  79961357.24  88862497.06 109768839.74 181987436.00

x <- subset(data$dmi_ingreso_total_mes, data$dmi_ingreso_total_mes <= 205163705.0)
hist(x)

dmi_ingreso_total_mes <- data$dmi_ingreso_total_mes

names(data)[49] # dh_val_pagos_d: Valor total de las trasacciones de salida por concepto de pagos otros

## [1] "dh_val_pagos_d"

summary(data$dh_val_pagos_d)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0   537640  3639828  4661394 46719192

hist(data$dh_val_pagos_d)

quantile(data$dh_val_pagos_d, seq(0, 1, by = 0.01))

##          0%          1%          2%          3%          4%          5% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##          6%          7%          8%          9%         10%         11% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         12%         13%         14%         15%         16%         17% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         18%         19%         20%         21%         22%         23% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         24%         25%         26%         27%         28%         29% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         30%         31%         32%         33%         34%         35% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         36%         37%         38%         39%         40%         41% 
##        0.00        0.00       75.34    42343.00    56773.00    69144.00 
##         42%         43%         44%         45%         46%         47% 
##    91390.00   151658.00   159690.00   165118.55   178014.00   179314.00 
##         48%         49%         50%         51%         52%         53% 
##   306936.36   318125.77   537640.00   678424.39   806356.20  1059441.20 
##         54%         55%         56%         57%         58%         59% 
##  1212291.94  1330827.25  1560711.00  1662900.00  1823001.70  1999493.24 
##         60%         61%         62%         63%         64%         65% 
##  2085516.80  2220129.00  2315083.28  2490672.54  2778419.00  3040516.20 
##         66%         67%         68%         69%         70%         71% 
##  3277671.00  3344574.00  3409709.20  3536084.43  3600516.00  3766572.79 
##         72%         73%         74%         75%         76%         77% 
##  3900425.24  4283458.20  4490363.42  4661393.75  4973548.56  5176261.00 
##         78%         79%         80%         81%         82%         83% 
##  5271188.86  5399363.49  5707806.00  6166031.00  6853466.00  7095437.00 
##         84%         85%         86%         87%         88%         89% 
##  7265797.40  7453799.00  7733374.12  8210036.00  9130377.00  9750499.50 
##         90%         91%         92%         93%         94%         95% 
## 10373706.00 10980804.97 12745205.00 13360645.10 13522747.00 14262631.00 
##         96%         97%         98%         99%        100% 
## 15040059.56 21305241.00 27212693.58 36549357.01 46719192.00

x <- subset(data$dh_val_pagos_d, data$dh_val_pagos_d <= 50357360.00)
hist(x)

dh_val_pagos_d <- data$dh_val_pagos_d

names(data)[50] # pc_gasto_familiar: Valor Gasto de familiar del cliente

## [1] "pc_gasto_familiar"

summary(data$pc_gasto_familiar)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0   998131  2284566  3596369  4568750 21888063

quantile(data$pc_gasto_familiar, seq(0,1,by=0.01))

##          0%          1%          2%          3%          4%          5% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##          6%          7%          8%          9%         10%         11% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         12%         13%         14%         15%         16%         17% 
##        0.00        0.00    63033.82   350000.00   491062.78   595099.36 
##         18%         19%         20%         21%         22%         23% 
##   665194.70   765000.00   829934.46   844698.40   864358.37   905320.38 
##         24%         25%         26%         27%         28%         29% 
##   943500.00   998130.70  1053864.93  1086803.67  1107646.36  1170000.00 
##         30%         31%         32%         33%         34%         35% 
##  1235790.67  1307972.22  1401226.17  1435756.50  1449533.38  1564995.31 
##         36%         37%         38%         39%         40%         41% 
##  1584240.00  1629589.42  1630082.22  1637151.00  1654248.75  1680000.00 
##         42%         43%         44%         45%         46%         47% 
##  1710290.10  1732919.27  1773123.19  1800533.81  1852381.20  2042659.37 
##         48%         49%         50%         51%         52%         53% 
##  2118915.12  2232519.95  2284565.98  2314858.89  2397248.78  2446335.13 
##         54%         55%         56%         57%         58%         59% 
##  2510217.45  2530766.61  2568167.58  2705762.50  2708752.54  2759275.31 
##         60%         61%         62%         63%         64%         65% 
##  2778750.00  2853154.49  2990963.55  3240028.70  3262926.88  3304375.00 
##         66%         67%         68%         69%         70%         71% 
##  3400000.00  3557347.50  3738034.50  3850755.50  3981578.82  4232305.25 
##         72%         73%         74%         75%         76%         77% 
##  4258569.80  4349935.73  4473926.14  4568750.00  4635750.00  4665646.00 
##         78%         79%         80%         81%         82%         83% 
##  5079723.23  5154531.56  5214205.69  5296256.50  5296256.50  5660150.00 
##         84%         85%         86%         87%         88%         89% 
##  5976224.40  6430000.00  7311417.07  7509873.84  8004993.21  8217756.65 
##         90%         91%         92%         93%         94%         95% 
##  9000354.71  9106196.62  9403175.00 10619436.29 12488403.37 13932843.75 
##         96%         97%         98%         99%        100% 
## 14228463.97 15245233.01 19531282.01 19549158.17 21888062.75

x <- subset(data$pc_gasto_familiar, data$pc_gasto_familiar <= 26582180.19)
hist(x)

pc_gasto_familiar <- data$pc_gasto_familiar

names(data)[51] # pc_cuotas_pagadas: Valor Cuotas de pagadas del cliente

## [1] "pc_cuotas_pagadas"

summary(data$pc_cuotas_pagadas)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0  343148  913284 1373781 7184545

hist(data$pc_cuotas_pagadas)

quantile(data$pc_cuotas_pagadas, seq(0, 1, by = 0.01))

##        0%        1%        2%        3%        4%        5%        6%        7% 
##       0.0       0.0       0.0       0.0       0.0       0.0       0.0       0.0 
##        8%        9%       10%       11%       12%       13%       14%       15% 
##       0.0       0.0       0.0       0.0       0.0       0.0       0.0       0.0 
##       16%       17%       18%       19%       20%       21%       22%       23% 
##       0.0       0.0       0.0       0.0       0.0       0.0       0.0       0.0 
##       24%       25%       26%       27%       28%       29%       30%       31% 
##       0.0       0.0       0.0       0.0     130.0   62150.0   87000.0  116000.0 
##       32%       33%       34%       35%       36%       37%       38%       39% 
##  143028.0  145000.0  179265.8  213500.0  220000.0  242425.2  258000.0  258000.0 
##       40%       41%       42%       43%       44%       45%       46%       47% 
##  258000.0  258000.0  258000.0  275855.6  280000.0  286810.0  294600.0  300000.0 
##       48%       49%       50%       51%       52%       53%       54%       55% 
##  332317.2  333200.0  343148.2  364200.7  384200.0  384698.3  400000.0  429782.6 
##       56%       57%       58%       59%       60%       61%       62%       63% 
##  473000.0  490000.0  504164.1  516000.0  543636.1  543636.1  646224.7  648815.1 
##       64%       65%       66%       67%       68%       69%       70%       71% 
##  676982.9  806622.1  847182.0  847182.0  855700.0  860000.0  874639.5  928418.7 
##       72%       73%       74%       75%       76%       77%       78%       79% 
## 1065817.2 1120496.9 1354695.3 1373781.0 1484196.6 1503863.1 1730412.1 1737940.1 
##       80%       81%       82%       83%       84%       85%       86%       87% 
## 1769838.0 1808000.0 1840035.3 1855943.0 1885749.1 1891250.5 2117686.3 2132206.4 
##       88%       89%       90%       91%       92%       93%       94%       95% 
## 2283470.0 2451854.4 2492565.3 2493312.4 2865653.8 2983499.5 3017035.4 3234839.9 
##       96%       97%       98%       99%      100% 
## 3303345.8 4616921.3 5167146.6 7155587.2 7184544.5

x <- subset(data$pc_cuotas_pagadas, data$pc_cuotas_pagadas <= 8156850.0)
hist(x)

pc_cuotas_pagadas <- data$pc_cuotas_pagadas

names(data)[52] # cpc_avg_proc_deuda: Promedio del porcentaje de las deudas de un cliente en el mes anterior (Porcentaje : Razón entre el saldo y el valor incial)

## [1] "cpc_avg_proc_deuda"

summary(data$cpc_avg_proc_deuda)

##       Min.    1st Qu.     Median       Mean    3rd Qu.       Max. 
## -28.182000   0.000000   0.003075   0.136219   0.799321   1.224909

hist(data$cpc_avg_proc_deuda)

quantile(data$cpc_avg_proc_deuda, seq(0,1,by=0.01))

##            0%            1%            2%            3%            4% 
## -28.182000000  -0.021213000  -0.003272055   0.000000000   0.000000000 
##            5%            6%            7%            8%            9% 
##   0.000000000   0.000000000   0.000000000   0.000000000   0.000000000 
##           10%           11%           12%           13%           14% 
##   0.000000000   0.000000000   0.000000000   0.000000000   0.000000000 
##           15%           16%           17%           18%           19% 
##   0.000000000   0.000000000   0.000000000   0.000000000   0.000000000 
##           20%           21%           22%           23%           24% 
##   0.000000000   0.000000000   0.000000000   0.000000000   0.000000000 
##           25%           26%           27%           28%           29% 
##   0.000000000   0.000000000   0.000000000   0.000000000   0.000000000 
##           30%           31%           32%           33%           34% 
##   0.000000000   0.000000000   0.000000000   0.000000000   0.000000000 
##           35%           36%           37%           38%           39% 
##   0.000000000   0.000000000   0.000000000   0.000000000   0.000000000 
##           40%           41%           42%           43%           44% 
##   0.000000000   0.000000000   0.000000000   0.000000000   0.000000000 
##           45%           46%           47%           48%           49% 
##   0.000000000   0.002474419   0.002631783   0.002631783   0.002631783 
##           50%           51%           52%           53%           54% 
##   0.003075038   0.018633346   0.059156000   0.096896600   0.115248053 
##           55%           56%           57%           58%           59% 
##   0.168734000   0.215616247   0.246319000   0.277754880   0.312850843 
##           60%           61%           62%           63%           64% 
##   0.362930164   0.402114125   0.427537092   0.442012067   0.478433000 
##           65%           66%           67%           68%           69% 
##   0.545081738   0.582370425   0.594232417   0.640208326   0.684193917 
##           70%           71%           72%           73%           74% 
##   0.700694736   0.746537701   0.757732590   0.776391909   0.784705667 
##           75%           76%           77%           78%           79% 
##   0.799320623   0.824395833   0.830392909   0.834281000   0.842780631 
##           80%           81%           82%           83%           84% 
##   0.853675575   0.858839117   0.868951947   0.895844211   0.929755381 
##           85%           86%           87%           88%           89% 
##   0.962535958   0.966394034   0.973559686   0.980652725   0.990590504 
##           90%           91%           92%           93%           94% 
##   1.017928827   1.045238187   1.072557625   1.085973066   1.098219571 
##           95%           96%           97%           98%           99% 
##   1.099494833   1.111387925   1.119784114   1.149961810   1.193426598 
##          100% 
##   1.224909400

names(data)[53] # cpc_sum_proc_deuda: Suma del porcentaje de las deudas de un cliente en el mes anterior (Porcentaje : Razón entre el saldo y el valor incial)

## [1] "cpc_sum_proc_deuda"

summary(data$cpc_sum_proc_deuda)

##       Min.    1st Qu.     Median       Mean    3rd Qu.       Max. 
## -28.182000   0.000000   0.003075   0.136219   0.799321   1.224909

hist(data$cpc_sum_proc_deuda)

names(data)[54] # dc_porc_prod_sin_mora: Porcentaje de productos sin mora en todo el sistema

## [1] "dc_porc_prod_sin_mora"

summary(data$dc_porc_prod_sin_mora)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
## 0.000000 0.000000 0.000000 0.005543 0.000000 1.000000

hist(data$dc_porc_prod_sin_mora)

quantile(data$dc_porc_prod_sin_mora, seq(0,1,by=0.01))

##   0%   1%   2%   3%   4%   5%   6%   7%   8%   9%  10%  11%  12%  13%  14%  15% 
##    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0 
##  16%  17%  18%  19%  20%  21%  22%  23%  24%  25%  26%  27%  28%  29%  30%  31% 
##    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0 
##  32%  33%  34%  35%  36%  37%  38%  39%  40%  41%  42%  43%  44%  45%  46%  47% 
##    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0 
##  48%  49%  50%  51%  52%  53%  54%  55%  56%  57%  58%  59%  60%  61%  62%  63% 
##    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0 
##  64%  65%  66%  67%  68%  69%  70%  71%  72%  73%  74%  75%  76%  77%  78%  79% 
##    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0 
##  80%  81%  82%  83%  84%  85%  86%  87%  88%  89%  90%  91%  92%  93%  94%  95% 
##    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0    0 
##  96%  97%  98%  99% 100% 
##    0    0    0    0    1

names(data)[55] # pc_ingreso_rutina_con_techo: Ingreso por rutina una vez aplicada los techos por segmento

## [1] "pc_ingreso_rutina_con_techo"

summary(data$pc_ingreso_rutina_con_techo)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0  3550276  8811787 15065481 20470286 87552251

hist(data$pc_ingreso_rutina_con_techo)

pc_ingreso_rutina_con_techo <- data$pc_ingreso_rutina_con_techo

names(data)[56] # pc_saldo_prom3_tdc_entidad: Saldo promedio de los últimos 3 meses de tarjeta de crédito en el banco

## [1] "pc_saldo_prom3_tdc_entidad"

summary(data$pc_saldo_prom3_tdc_entidad)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0    5652       0 3070000

hist(data$pc_saldo_prom3_tdc_entidad)

names(data)[57] # dh_cant_salidas: Cantidad de trasacciones de salida de dinero en un mes

## [1] "dh_cant_salidas"

summary(data$dh_cant_salidas)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    0.00    0.00   19.00   28.66   49.00  184.00

hist(data$dh_cant_salidas)

dh_cant_salidas <- data$dh_cant_salidas

names(data)[58] # dh_min_dia_pagos_d: Primer dia en el que realizó algún credito en el mes anterior

## [1] "dh_min_dia_pagos_d"

summary(data$dh_min_dia_pagos_d)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   0.000   0.000   1.000   3.171   3.000  31.000

hist(data$dh_min_dia_pagos_d)

dh_min_dia_pagos_d <- data$dh_min_dia_pagos_d

names(data)[59] # pc_ingreso_por_rutina: Ingreso por rutina

## [1] "pc_ingreso_por_rutina"

summary(data$pc_ingreso_por_rutina)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0  3550276  8811787 15083531 20470286 87552251

hist(data$pc_ingreso_por_rutina)

pc_ingreso_por_rutina <- data$pc_ingreso_por_rutina

names(data)[60] # dh_min_dia_pago_tarj_d: Primer dia en el que pago la tarjeta de credito en el mes anterior

## [1] "dh_min_dia_pago_tarj_d"

hist(data$dh_min_dia_pago_tarj_d)

names(data)[61] # cp_nro_cuota: Numero de cuota pactadas por producto

## [1] "cp_nro_cuota"

summary(data$cp_nro_cuota)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[62] # dmi_egreso_total_mes: Egreso total del mes anterior

## [1] "dmi_egreso_total_mes"

summary(data$dmi_egreso_total_mes)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0   6710852  18203268  24720956 191314996

hist(data$dmi_egreso_total_mes)

quantile(data$dmi_egreso_total_mes, seq(0,1,by=0.01))

##           0%           1%           2%           3%           4%           5% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##           6%           7%           8%           9%          10%          11% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          12%          13%          14%          15%          16%          17% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          18%          19%          20%          21%          22%          23% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          24%          25%          26%          27%          28%          29% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          30%          31%          32%          33%          34%          35% 
##         0.00         0.00         0.00         0.00         0.00     11399.35 
##          36%          37%          38%          39%          40%          41% 
##     65750.00    144062.00    476365.00    857438.55   1373997.80   2032524.46 
##          42%          43%          44%          45%          46%          47% 
##   2425098.10   3147436.28   3637537.36   4297695.00   4985629.00   5242864.00 
##          48%          49%          50%          51%          52%          53% 
##   5768836.60   6471482.00   6710852.00   7234899.00   7941029.00   8321595.04 
##          54%          55%          56%          57%          58%          59% 
##   9066380.48   9451729.00  10147091.00  11474139.00  11927600.00  12472455.70 
##          60%          61%          62%          63%          64%          65% 
##  13236143.00  13635595.00  14447329.22  15199359.70  16050052.32  16698889.00 
##          66%          67%          68%          69%          70%          71% 
##  17675826.00  18687460.99  19390695.00  20624901.00  21419482.00  21608609.00 
##          72%          73%          74%          75%          76%          77% 
##  22283315.00  23681242.17  23844136.20  24720956.00  25905897.48  27633120.00 
##          78%          79%          80%          81%          82%          83% 
##  29004586.56  30127755.34  31398939.00  31492755.00  32045289.72  35278365.00 
##          84%          85%          86%          87%          88%          89% 
##  40680145.00  43554385.50  45389761.66  48031045.00  52135165.00  56926454.85 
##          90%          91%          92%          93%          94%          95% 
##  60640219.00  64217571.00  66239075.00  68137836.00  71343296.66  78399975.45 
##          96%          97%          98%          99%         100% 
##  80045627.00  80352549.82  89004556.92 111049163.72 191314996.00

x <- subset(data$dmi_egreso_total_mes, data$dmi_egreso_total_mes <= 209829245.0)
hist(x)

dmi_egreso_total_mes <- data$dmi_egreso_total_mes

names(data)[63] # cp_valor_inicial: valor inicial de la obligacion por producto

## [1] "cp_valor_inicial"

summary(data$cp_valor_inicial)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[64] # dh_max_dia_otros_d: Ultimo dia en el que realizó algún débito clasificado como "otros" en el mes anterior

## [1] "dh_max_dia_otros_d"

hist(data$dh_max_dia_otros_d)

x <- ifelse(data$dh_max_dia_otros_d <= 15, 1, 0)
table(x)

## x
##   0   1 
## 387 515

flax_maxDiaOtrosD_1quincena <- x

names(data)[65] # p_cuota_sobre_saldo: Valor de la cuota sobre saldo por producto

## [1] "cp_cuota_sobre_saldo"

summary(data$cp_cuota_sobre_saldo)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[66] # dc_sum_valor_inicial: Suma valores iniciales de obligaciones en todo el sistema financiero

## [1] "dc_sum_valor_inicial"

summary(data$dc_sum_valor_inicial)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0         0    501971         0 310276000

hist(data$dc_sum_valor_inicial)

names(data)[67] # dh_cant_pago_tarj_d: Cantidad de trasacciones de salida por concepto de pago de tarjeta de credito

## [1] "dh_cant_pago_tarj_d"

summary(data$dh_cant_pago_tarj_d)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##  0.0000  0.0000  0.0000  0.9678  1.0000 33.0000

hist(data$dh_cant_pago_tarj_d)

names(data)[68] # dh_max_dia_pagos_d: Ultimo dia en el que realizó algún credito en el mes anterior

## [1] "dh_max_dia_pagos_d"

hist(data$dh_max_dia_pagos_d)

x <- ifelse(data$dh_max_dia_pagos_d <= 15, 1, 0)
table(x)

## x
##   0   1 
## 505 397

flag_maxDiaPagos_d_1quincena <- x

names(data)[69] # cp_saldo_sobre_inicial: valor del saldo sobre inicial por producto mes anterior

## [1] "cp_saldo_sobre_inicial"

summary(data$cp_saldo_sobre_inicial)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[70] # pc_mediana_nom3: Mediana de los últimos 3 pagos nómina  para cálculo de retanqueo libranza

## [1] "pc_mediana_nom3"

summary(data$pc_mediana_nom3)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0   50084       0 4287978

hist(data$pc_mediana_nom3)

names(data)[71] # cp_esta_cuota_otro: Es el estado de la cuota otro

## [1] "cp_esta_cuota_otro"

summary(data$cp_esta_cuota_otro)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[72] # dh_max_dia_retiros_d: Ultimo dia en el que realizó reitro en el mes anterior

## [1] "dh_max_dia_retiros_d"

hist(data$dh_max_dia_retiros_d)

# variable incorrecta, el cuartil 3 es 30 y un mes tiene hasta 31 días

names(data)[73] # dh_avg_dia_entradas: Dia promedio del mes en el recibe las entradas de dinero

## [1] "dh_avg_dia_entradas"

summary(data$dh_avg_dia_entradas)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   0.000   0.000   4.585   4.457   6.557  31.000

hist(data$dh_avg_dia_entradas)

names(data)[74] # dh_avg_dia_pagos_d: Dia del mes promedio en el que hace las salidas de dinero por pagos

## [1] "dh_avg_dia_pagos_d"

x <- ifelse(data$dh_avg_dia_pagos_d <= 10, 1, 0)
table(x)

## x
##   0   1 
##   1 901

names(data)[75] # dh_val_salidas: Valor total de las trasacciones de salida en un mes

## [1] "dh_val_salidas"

summary(data$dh_val_salidas)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0   6710852  18220966  24720956 191314996

hist(data$dh_val_salidas)

quantile(data$dh_val_salidas, seq(0,1,by=0.01))

##           0%           1%           2%           3%           4%           5% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##           6%           7%           8%           9%          10%          11% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          12%          13%          14%          15%          16%          17% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          18%          19%          20%          21%          22%          23% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          24%          25%          26%          27%          28%          29% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          30%          31%          32%          33%          34%          35% 
##         0.00         0.00         0.00         0.00     11399.34     38171.30 
##          36%          37%          38%          39%          40%          41% 
##     97964.00    275655.54    524117.36   1168205.09   1781441.00   2117572.83 
##          42%          43%          44%          45%          46%          47% 
##   2514844.00   3349365.00   3817628.00   4412149.25   5013252.92   5242864.00 
##          48%          49%          50%          51%          52%          53% 
##   5768836.60   6471482.00   6710852.00   7234899.00   7941029.00   8321595.04 
##          54%          55%          56%          57%          58%          59% 
##   9066380.48   9451729.00  10147091.00  11474139.00  11927600.00  12472455.70 
##          60%          61%          62%          63%          64%          65% 
##  13236143.00  13635595.00  14447329.22  15199359.70  16050052.32  16698889.00 
##          66%          67%          68%          69%          70%          71% 
##  17675826.00  18687460.99  19390695.00  20624901.00  21419482.00  21608609.00 
##          72%          73%          74%          75%          76%          77% 
##  22283315.00  23681242.17  23844136.20  24720956.00  25905897.48  27633120.00 
##          78%          79%          80%          81%          82%          83% 
##  29004586.56  30127755.34  31398939.00  31492755.00  32045289.72  35278365.00 
##          84%          85%          86%          87%          88%          89% 
##  40680145.00  43554385.50  45389761.66  48031045.00  52135165.00  56926454.85 
##          90%          91%          92%          93%          94%          95% 
##  60640219.00  64217571.00  66239075.00  68137836.00  71343296.66  78399975.45 
##          96%          97%          98%          99%         100% 
##  80045627.00  80352549.82  89004556.92 111049163.72 191314996.00

x <- subset(data$dh_val_salidas, data$dh_val_salidas <= 217011100)
hist(x)

dh_val_salidas <- data$dh_val_salidas

names(data)[76] # dc_sum_valor_cuota: Valor total de las trasacciones de salida en un mes

## [1] "dc_sum_valor_cuota"

summary(data$dc_sum_valor_cuota)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0        0    33407        0 13360000

hist(data$dc_sum_valor_cuota)

names(data)[77] # dh_min_dia_tras_d: Primer dia en el que realizó pago de traslado en el mes anterior

## [1] "dh_min_dia_tras_d"

hist(data$dh_min_dia_tras_d)

x <- ifelse(data$dh_min_dia_tras_d <= 10, 1, 0)
table(x)

## x
##   0   1 
##  99 803

names(data)[78] # cp_porc_valorcuot_ing: Relación entre el valor de la cuota sobre los ingresos por producto

## [1] "cp_porc_valorcuot_ing"

summary(data$cp_porc_valorcuot_ing)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[79] # pc_ind_ajustado: Ingreso neto disponible  del cliente ajustado

## [1] "pc_ind_ajustado"

summary(data$pc_ind_ajustado)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
## -4115281   994836  3591609  6805845  9008932 46552980

hist(data$pc_ind_ajustado)

quantile(data$pc_ind_ajustado, seq(0,1,by=0.01))

##          0%          1%          2%          3%          4%          5% 
## -4115280.80 -1826179.99  -289405.70   -96562.97        0.00        0.00 
##          6%          7%          8%          9%         10%         11% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         12%         13%         14%         15%         16%         17% 
##        0.00        0.00        0.00        0.00        0.00    48003.75 
##         18%         19%         20%         21%         22%         23% 
##   109162.44   190428.75   246883.42   538385.69   648167.69   799010.24 
##         24%         25%         26%         27%         28%         29% 
##   859590.00   994835.77  1157424.69  1289365.08  1350754.47  1404675.00 
##         30%         31%         32%         33%         34%         35% 
##  1480053.89  1587621.36  1587621.36  1650235.75  1822835.97  1879208.30 
##         36%         37%         38%         39%         40%         41% 
##  1948169.40  2003917.50  2154537.00  2518881.75  2677500.00  2822090.15 
##         42%         43%         44%         45%         46%         47% 
##  2911943.92  3091390.95  3190972.67  3190972.67  3241722.38  3241722.38 
##         48%         49%         50%         51%         52%         53% 
##  3413106.65  3505538.87  3591609.21  3591609.21  3872384.69  4317499.00 
##         54%         55%         56%         57%         58%         59% 
##  4357215.99  4474653.75  4668131.63  4848009.69  4974054.87  5360090.10 
##         60%         61%         62%         63%         64%         65% 
##  5741412.68  5741412.68  5777751.94  5859048.42  5895596.49  6025889.71 
##         66%         67%         68%         69%         70%         71% 
##  6098398.45  6318411.44  6806462.97  7365699.76  7610175.00  7897292.18 
##         72%         73%         74%         75%         76%         77% 
##  8258805.03  8432540.03  8888342.46  9008932.31 10100269.49 10443540.70 
##         78%         79%         80%         81%         82%         83% 
## 11475000.00 11744826.48 11744826.48 11854694.27 12430296.87 13127593.73 
##         84%         85%         86%         87%         88%         89% 
## 13154771.40 14039078.44 14122015.13 15690171.27 15790025.89 18134416.04 
##         90%         91%         92%         93%         94%         95% 
## 18941384.29 21621182.46 22345186.29 23339886.01 23429767.22 24527785.16 
##         96%         97%         98%         99%        100% 
## 26203986.86 30567968.50 30946818.46 36547579.52 46552980.18

x <- subset(data$pc_ind_ajustado, data$pc_ind_ajustado <= 68107989.8)
hist(x)

names(data)[80] # dh_val_retiros_d: Valor total de las trasacciones de salida por concepto de retiros en un mes

## [1] "dh_val_retiros_d"

summary(data$dh_val_retiros_d)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##         0         0   1048890   8526310  10895497 164054718

hist(data$dh_val_retiros_d)

names(data)[81] # pc_tiem_lt_prod_abie_total: Tiempo último producto abierto

## [1] "pc_tiem_lt_prod_abie_total"

summary(data$pc_tiem_lt_prod_abie_total)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
## 0.00000 0.00000 0.00000 0.03104 0.00000 9.00000

hist(data$pc_tiem_lt_prod_abie_total)

names(data)[82] # marca_info_cifin_decode: Marca Cifin(Consultado, no consultado, no encontrado, etc)

## [1] "marca_info_cifin_decode"

table(data$marca_info_cifin_decode)

## 
##   0   1   2 
## 645   5 252

x <- ifelse(data$marca_info_cifin_decode == 0, 1, 0)
table(x)

## x
##   0   1 
## 257 645

flag_encontrado_cifin <- x

names(data)[83] # dh_max_dia_pago_tarj_d: Ultimo dia en el que pago la tarjeta de credito en el mes anterior

## [1] "dh_max_dia_pago_tarj_d"

hist(data$dh_max_dia_pago_tarj_d)

names(data)[84] # dc_valobli_ing: Suma del valor inicial de las obligaciones en el sistema financiero sobre los ingresos

## [1] "dc_valobli_ing"

summary(data$dc_valobli_ing)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##  0.00000  0.00000  0.00000  0.03364  0.00000 23.86188

hist(data$dc_valobli_ing)

names(data)[85] # pc_cantidad_tdc_entidad: Cantidad tarjetas de crédito en el banco

## [1] "pc_cantidad_tdc_entidad"

table(data$pc_cantidad_tdc_entidad)

## 
##   0   1   2 
## 899   2   1

names(data)[86] # dh_min_dia_otros_d: Primer dia en el que realizó algún débito clasificado como "otros" en el mes anterior

## [1] "dh_min_dia_otros_d"

hist(data$dh_min_dia_otros_d)

x <- ifelse(data$dh_min_dia_otros_d <= 15, 1, 0)
table(x)

## x
##   0   1 
##  65 837

names(data)[87] # dc_cant_obligaciones: Cantidad de obligaciones

## [1] "dc_cant_obligaciones"

table(data$dc_cant_obligaciones)

## 
##   0   3   8  10  15 
## 897   1   2   1   1

summary(data$dc_cant_obligaciones)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##  0.00000  0.00000  0.00000  0.04878  0.00000 15.00000

hist(data$dc_cant_obligaciones)

x <- ifelse(data$dc_cant_obligaciones >= 1, 1, 0)
table(x)

## x
##   0   1 
## 897   5

names(data)[88] # cpc_sum_nro_cuota: Suma de los número de cuotas de todas las obligaciones del cliente

## [1] "cpc_sum_nro_cuota"

summary(data$cpc_sum_nro_cuota)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##     0.0     0.0    25.5    93.3   134.0   778.0

names(data)[89] # cpc_avg_nro_cuota: Promedio del número de cuotas entre todos los productos del cliente

## [1] "cpc_avg_nro_cuota"

summary(data$cpc_avg_nro_cuota)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    0.00    0.00   20.33   39.96   66.38  305.50

hist(data$cpc_avg_nro_cuota)

cpc_avg_nro_cuota <- data$cpc_avg_nro_cuota

names(data)[90] # cpc_max_nro_cuota: Número de cuotas máximo de un cliente entre todos sus productos

## [1] "cpc_max_nro_cuota"

names(data)[91] # cp_saldo: Valor del saldo del mes anterior por producto

## [1] "cp_saldo"

summary(data$cp_saldo)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[92] # cp_cuota_sobre_inicial: valor de la cuota sobre el valor desembolsado incial por producto

## [1] "cp_cuota_sobre_inicial"

summary(data$cp_cuota_sobre_inicial)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[93] # cpc_sum_saldo: Suma del saldo de todos los productos activos del cliente

## [1] "cpc_sum_saldo"

summary(data$cpc_sum_saldo)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##    -70455         0     33950  12216024   9920492 146717108

hist(data$cpc_sum_saldo)

quantile(data$cpc_sum_saldo, seq(0,1,by=0.01))

##           0%           1%           2%           3%           4%           5% 
##    -70455.00    -65826.00    -19656.58         0.00         0.00         0.00 
##           6%           7%           8%           9%          10%          11% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          12%          13%          14%          15%          16%          17% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          18%          19%          20%          21%          22%          23% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          24%          25%          26%          27%          28%          29% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          30%          31%          32%          33%          34%          35% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          36%          37%          38%          39%          40%          41% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          42%          43%          44%          45%          46%          47% 
##         0.00         0.00         0.00         0.00     24490.00     31920.00 
##          48%          49%          50%          51%          52%          53% 
##     33950.00     33950.00     33950.00    237879.40    472658.20    738957.00 
##          54%          55%          56%          57%          58%          59% 
##    985207.00   1420982.55   1946855.76   2051116.00   2437914.82   2604606.00 
##          60%          61%          62%          63%          64%          65% 
##   2704777.80   3033503.54   3199161.00   3792700.64   4372629.00   4874872.00 
##          66%          67%          68%          69%          70%          71% 
##   5131967.56   5649048.00   7246359.24   7536951.00   7706847.00   8010945.00 
##          72%          73%          74%          75%          76%          77% 
##   8457241.00   8746891.55   9134322.00   9920492.50  11517312.00  12052637.97 
##          78%          79%          80%          81%          82%          83% 
##  12249094.00  14379750.00  15182781.20  16772655.45  17733736.74  18370147.00 
##          84%          85%          86%          87%          88%          89% 
##  19785500.00  23100863.00  23505359.00  26303831.91  27975363.72  29763054.02 
##          90%          91%          92%          93%          94%          95% 
##  33065306.00  36919522.00  47807424.63  73378700.49  78844063.50  80180499.99 
##          96%          97%          98%          99%         100% 
##  83182148.00  84670677.73 112508218.03 128558087.07 146717108.20

x <- subset(data$cpc_sum_saldo, data$cpc_sum_saldo <= 220905770.00)
hist(x)

cpc_sum_saldo <- data$cpc_sum_saldo

names(data)[94] # cp_porc_saldo_ing: valor del saldo sobre ingreso por producto

## [1] "cp_porc_saldo_ing"

summary(data$cp_porc_saldo_ing)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[95] # gsm_mejor_gestion: Mejor gestión realizada en el mes anterior

## [1] "gsm_mejor_gestion"

table(data$gsm_mejor_gestion)

## 
##   0  16 
## 901   1

names(data)[96] # dh_min_dia_nomina_c

## [1] "dh_min_dia_nomina_c"

names(data)[97] # dh_max_dia_nomina_c: Ultimo dia en el que recibió pago de nomina en el mes anterior

## [1] "dh_max_dia_nomina_c"

hist(data$dh_max_dia_nomina_c)

names(data)[98] # cp_valor_cuota: Valor de cuota por producto

## [1] "cp_valor_cuota"

summary(data$cp_valor_cuota)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0       0       0       0

names(data)[99] # cpc_nro_cuota_tdc: Suma de los número de cuotas de todas las tarjetas de crédito del cliente

## [1] "cpc_nro_cuota_tdc"

names(data)[100] # gsm_prom_dias_gest: Promedio de los días en que se realizaron gestiones en el mes anterior

## [1] "gsm_prom_dias_gest"

names(data)[101] # pc_cuota_no_rot_ent: Cuota de productos no rotativos en el banco

## [1] "pc_cuota_no_rot_ent"

summary(data$pc_cuota_no_rot_ent)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0    1527       0 1377000

hist(data$pc_cuota_no_rot_ent)

names(data)[102] # dh_val_nomina_c: Valor total de las trasacciones de entrada por concepto de nomina en un mes

## [1] "dh_val_nomina_c"

summary(data$dh_val_nomina_c)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0   60297       0 7256848

hist(data$dh_val_nomina_c)

quantile(data$dh_val_nomina_c, seq(0,1,by=0.01))

##      0%      1%      2%      3%      4%      5%      6%      7%      8%      9% 
##       0       0       0       0       0       0       0       0       0       0 
##     10%     11%     12%     13%     14%     15%     16%     17%     18%     19% 
##       0       0       0       0       0       0       0       0       0       0 
##     20%     21%     22%     23%     24%     25%     26%     27%     28%     29% 
##       0       0       0       0       0       0       0       0       0       0 
##     30%     31%     32%     33%     34%     35%     36%     37%     38%     39% 
##       0       0       0       0       0       0       0       0       0       0 
##     40%     41%     42%     43%     44%     45%     46%     47%     48%     49% 
##       0       0       0       0       0       0       0       0       0       0 
##     50%     51%     52%     53%     54%     55%     56%     57%     58%     59% 
##       0       0       0       0       0       0       0       0       0       0 
##     60%     61%     62%     63%     64%     65%     66%     67%     68%     69% 
##       0       0       0       0       0       0       0       0       0       0 
##     70%     71%     72%     73%     74%     75%     76%     77%     78%     79% 
##       0       0       0       0       0       0       0       0       0       0 
##     80%     81%     82%     83%     84%     85%     86%     87%     88%     89% 
##       0       0       0       0       0       0       0       0       0       0 
##     90%     91%     92%     93%     94%     95%     96%     97%     98%     99% 
##       0       0       0       0       0       0       0       0       0 3227685 
##    100% 
## 7256848

names(data)[103] # banca_completa: Si el cliente pertenece al segmento banca completa o no

## [1] "banca_completa"

table(data$banca_completa)

## 
##   0 
## 902

names(data)[104] # cpc_saldo_sobre_ing: saldo sobre ingreso por cliente

## [1] "cpc_saldo_sobre_ing"

summary(data$cpc_saldo_sobre_ing)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
## -0.01392  0.00000  0.00392  0.75912  0.43980 11.63406

hist(data$cpc_saldo_sobre_ing)

quantile(data$cpc_saldo_sobre_ing, seq(0,1,by=0.01))

##            0%            1%            2%            3%            4% 
## -0.0139196450 -0.0026716620 -0.0004434407  0.0000000000  0.0000000000 
##            5%            6%            7%            8%            9% 
##  0.0000000000  0.0000000000  0.0000000000  0.0000000000  0.0000000000 
##           10%           11%           12%           13%           14% 
##  0.0000000000  0.0000000000  0.0000000000  0.0000000000  0.0000000000 
##           15%           16%           17%           18%           19% 
##  0.0000000000  0.0000000000  0.0000000000  0.0000000000  0.0000000000 
##           20%           21%           22%           23%           24% 
##  0.0000000000  0.0000000000  0.0000000000  0.0000000000  0.0000000000 
##           25%           26%           27%           28%           29% 
##  0.0000000000  0.0000000000  0.0000000000  0.0000000000  0.0000000000 
##           30%           31%           32%           33%           34% 
##  0.0000000000  0.0000000000  0.0000000000  0.0000000000  0.0000000000 
##           35%           36%           37%           38%           39% 
##  0.0000000000  0.0000000000  0.0000000000  0.0000000000  0.0000000000 
##           40%           41%           42%           43%           44% 
##  0.0000000000  0.0000000000  0.0000000000  0.0000000000  0.0000000000 
##           45%           46%           47%           48%           49% 
##  0.0000000000  0.0006086835  0.0036859120  0.0039203230  0.0039203230 
##           50%           51%           52%           53%           54% 
##  0.0039203230  0.0091222870  0.0160144150  0.0241167220  0.0342690350 
##           55%           56%           57%           58%           59% 
##  0.0516975909  0.0762671575  0.0871885689  0.0959601992  0.0996937647 
##           60%           61%           62%           63%           64% 
##  0.1100301990  0.1167548170  0.1181152130  0.1414183478  0.1527656520 
##           65%           66%           67%           68%           69% 
##  0.2038563155  0.2406806960  0.2590162859  0.2652612750  0.2722557190 
##           70%           71%           72%           73%           74% 
##  0.3197385828  0.3433782008  0.3739022130  0.3850143830  0.4290230500 
##           75%           76%           77%           78%           79% 
##  0.4397979330  0.5008516424  0.5767091904  0.6159902625  0.7295992450 
##           80%           81%           82%           83%           84% 
##  0.8236684390  0.8864274193  0.9934920512  1.0631846570  1.1306370686 
##           85%           86%           87%           88%           89% 
##  1.1725284910  1.2668796990  1.4941067817  1.8447025345  1.9151466040 
##           90%           91%           92%           93%           94% 
##  2.1176921894  2.1883557560  2.2844911496  2.9200289999  5.4915779623 
##           95%           96%           97%           98%           99% 
##  5.8321611310  6.7831752090  7.1059412270  7.4480707058  8.7215061950 
##          100% 
## 11.6340582200

x <- subset(data$cpc_saldo_sobre_ing, data$cpc_saldo_sobre_ing <= 20.307294513)
hist(x)

cpc_saldo_sobre_ing <- data$cpc_saldo_sobre_ing

names(data)[105] # dh_min_dia_pago_cred_d: Primer dia en el que realizó pago de credito en el mes anterior

## [1] "dh_min_dia_pago_cred_d"

hist(data$dh_min_dia_pago_cred_d)

dh_min_dia_pago_cred_d <- data$dh_min_dia_pago_cred_d

names(data)[106] # cpc_saldo_tdc: Saldo total en tarjetas de crédito del cliente

## [1] "cpc_saldo_tdc"

summary(data$cpc_saldo_tdc)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##   -70455        0    33950  5068261  8010945 36919522

hist(data$cpc_saldo_tdc)

quantile(data$cpc_saldo_tdc, seq(0,1,by=0.01))

##          0%          1%          2%          3%          4%          5% 
##   -70455.00   -65826.00   -20000.00        0.00        0.00        0.00 
##          6%          7%          8%          9%         10%         11% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         12%         13%         14%         15%         16%         17% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         18%         19%         20%         21%         22%         23% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         24%         25%         26%         27%         28%         29% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         30%         31%         32%         33%         34%         35% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         36%         37%         38%         39%         40%         41% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         42%         43%         44%         45%         46%         47% 
##        0.00        0.00        0.00        0.00        0.00     9483.00 
##         48%         49%         50%         51%         52%         53% 
##    31920.00    33950.00    33950.00    33950.00    40620.92   242390.33 
##         54%         55%         56%         57%         58%         59% 
##   350295.68   517486.50   738957.00   985207.00  1425073.34  1883099.63 
##         60%         61%         62%         63%         64%         65% 
##  2025065.80  2296037.81  2585201.00  2642201.49  2905231.56  3156513.95 
##         66%         67%         68%         69%         70%         71% 
##  3494659.42  4090121.68  4507996.36  4949410.41  5262763.00  5664027.00 
##         72%         73%         74%         75%         76%         77% 
##  7246970.96  7536951.00  7706353.52  8010945.00  8318098.24  8616938.17 
##         78%         79%         80%         81%         82%         83% 
##  9134322.00  9296475.24 10437137.80 11181913.72 11687764.00 12248475.71 
##         84%         85%         86%         87%         88%         89% 
## 12852364.00 13440597.70 14740435.54 15089862.00 16685620.00 17175586.00 
##         90%         91%         92%         93%         94%         95% 
## 17905425.00 18549522.56 19785500.00 21713390.07 22926946.00 23100863.00 
##         96%         97%         98%         99%        100% 
## 24470898.84 26387876.00 28790737.34 31877765.63 36919522.00

x <- subset(data$cpc_saldo_tdc, data$cpc_saldo_tdc <= 43186883.96)
hist(x)

cpc_saldo_tdc <- data$cpc_saldo_tdc

names(data)[107] # pc_cuota_de_consumo: Cuota de crédito  de consumo reportada por CIFIN

## [1] "pc_cuota_de_consumo"

summary(data$pc_cuota_de_consumo)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##       0       0       0    5096       0 3356000

hist(data$pc_cuota_de_consumo)

names(data)[108] # dh_cant_tras_d: Cantidad de traslados de salida de dinero en un mes

## [1] "dh_cant_tras_d"

hist(data$dh_cant_tras_d)

quantile(data$dh_cant_tras_d, seq(0,1,by=0.005))

##   0.0%   0.5%   1.0%   1.5%   2.0%   2.5%   3.0%   3.5%   4.0%   4.5%   5.0% 
##  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000 
##   5.5%   6.0%   6.5%   7.0%   7.5%   8.0%   8.5%   9.0%   9.5%  10.0%  10.5% 
##  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000 
##  11.0%  11.5%  12.0%  12.5%  13.0%  13.5%  14.0%  14.5%  15.0%  15.5%  16.0% 
##  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000 
##  16.5%  17.0%  17.5%  18.0%  18.5%  19.0%  19.5%  20.0%  20.5%  21.0%  21.5% 
##  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000 
##  22.0%  22.5%  23.0%  23.5%  24.0%  24.5%  25.0%  25.5%  26.0%  26.5%  27.0% 
##  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000 
##  27.5%  28.0%  28.5%  29.0%  29.5%  30.0%  30.5%  31.0%  31.5%  32.0%  32.5% 
##  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000 
##  33.0%  33.5%  34.0%  34.5%  35.0%  35.5%  36.0%  36.5%  37.0%  37.5%  38.0% 
##  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000 
##  38.5%  39.0%  39.5%  40.0%  40.5%  41.0%  41.5%  42.0%  42.5%  43.0%  43.5% 
##  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000 
##  44.0%  44.5%  45.0%  45.5%  46.0%  46.5%  47.0%  47.5%  48.0%  48.5%  49.0% 
##  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000 
##  49.5%  50.0%  50.5%  51.0%  51.5%  52.0%  52.5%  53.0%  53.5%  54.0%  54.5% 
##  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000 
##  55.0%  55.5%  56.0%  56.5%  57.0%  57.5%  58.0%  58.5%  59.0%  59.5%  60.0% 
##  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000  0.000 
##  60.5%  61.0%  61.5%  62.0%  62.5%  63.0%  63.5%  64.0%  64.5%  65.0%  65.5% 
##  0.000  0.000  0.000  1.000  1.000  1.000  1.000  1.000  1.000  1.000  1.000 
##  66.0%  66.5%  67.0%  67.5%  68.0%  68.5%  69.0%  69.5%  70.0%  70.5%  71.0% 
##  1.000  1.000  1.000  1.000  1.000  1.000  1.000  1.000  2.000  2.000  2.000 
##  71.5%  72.0%  72.5%  73.0%  73.5%  74.0%  74.5%  75.0%  75.5%  76.0%  76.5% 
##  2.000  2.000  2.000  2.000  2.000  2.000  3.000  3.000  3.000  3.000  3.000 
##  77.0%  77.5%  78.0%  78.5%  79.0%  79.5%  80.0%  80.5%  81.0%  81.5%  82.0% 
##  3.000  3.000  4.000  4.000  4.000  4.000  4.000  4.000  4.000  5.000  5.000 
##  82.5%  83.0%  83.5%  84.0%  84.5%  85.0%  85.5%  86.0%  86.5%  87.0%  87.5% 
##  5.000  5.000  6.000  6.000  6.000  6.000  6.000  6.000  6.000  7.000  7.000 
##  88.0%  88.5%  89.0%  89.5%  90.0%  90.5%  91.0%  91.5%  92.0%  92.5%  93.0% 
##  7.000  8.000  8.000  9.000  9.000  9.000  9.000 10.000 10.000 10.000 10.000 
##  93.5%  94.0%  94.5%  95.0%  95.5%  96.0%  96.5%  97.0%  97.5%  98.0%  98.5% 
## 11.000 11.940 12.000 13.000 13.000 14.000 15.465 16.000 18.000 20.000 21.000 
##  99.0%  99.5% 100.0% 
## 21.000 21.495 30.000

x <- subset(data$dh_cant_tras_d, data$dh_cant_tras_d <= 31)
hist(x)

dh_cant_tras_d <- data$dh_cant_tras_d

names(data)[109] # dh_max_dia_comisio_d

## [1] "dh_max_dia_comisio_d"

names(data)[110] # cpc_avg_saldo: Promedio del saldo de las obligaciones del cliente en el mes anterior

## [1] "cpc_avg_saldo"

summary(data$cpc_avg_saldo)

##      Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
##    -70455         0     16975   4949118   5029954 103158078

hist(data$cpc_avg_saldo)

quantile(data$cpc_avg_saldo, seq(0,1,by=0.01))

##           0%           1%           2%           3%           4%           5% 
##    -70455.00    -65826.00    -19628.29         0.00         0.00         0.00 
##           6%           7%           8%           9%          10%          11% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          12%          13%          14%          15%          16%          17% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          18%          19%          20%          21%          22%          23% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          24%          25%          26%          27%          28%          29% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          30%          31%          32%          33%          34%          35% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          36%          37%          38%          39%          40%          41% 
##         0.00         0.00         0.00         0.00         0.00         0.00 
##          42%          43%          44%          45%          46%          47% 
##         0.00         0.00         0.00         0.00     14169.66     16975.00 
##          48%          49%          50%          51%          52%          53% 
##     16975.00     16975.00     16975.00    118939.70    305999.12    433154.00 
##          54%          55%          56%          57%          58%          59% 
##    601811.48    796185.15    993033.54   1175050.00   1302303.00   1493580.88 
##          60%          61%          62%          63%          64%          65% 
##   1694424.33   1897335.85   2052581.75   2283110.98   2580505.34   2724391.55 
##          66%          67%          68%          69%          70%          71% 
##   2921941.00   3211054.84   3662326.44   3909918.00   4267080.33   4304110.50 
##          72%          73%          74%          75%          76%          77% 
##   4535613.02   4567161.00   4864807.92   5029954.00   5649048.00   6121840.71 
##          78%          79%          80%          81%          82%          83% 
##   6553668.23   6977784.50   7230429.80   7465720.25   7623077.00   7910066.74 
##          84%          85%          86%          87%          88%          89% 
##   8316979.75   9229880.50  10935486.07  11550431.50  12813655.74  13193938.00 
##          90%          91%          92%          93%          94%          95% 
##  13839320.64  14111779.62  16685620.00  17892250.14  22239378.76  24469808.29 
##          96%          97%          98%          99%         100% 
##  26367451.45  28919972.77  42970063.02  55008435.80 103158078.10

x <- subset(data$cpc_avg_saldo, data$cpc_avg_saldo <= 106884393.50)
hist(x)

cpc_avg_saldo <- data$cpc_avg_saldo

names(data)[111] # dc_max_saldo_sf: máximo saldo en el sistema financiero (Solo Financiero)

## [1] "dc_max_saldo_sf"

summary(data$dc_max_saldo_sf)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0        0   167955        0 69454000

names(data)[112] # dh_val_pago_tarj_d: Valor total de las trasacciones de salida por concepto de pagos de tdc en un mes

## [1] "dh_val_pago_tarj_d"

summary(data$dh_val_pago_tarj_d)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0        0   555679    61884 10050161

hist(data$dh_val_pago_tarj_d)

quantile(data$dh_val_pago_tarj_d, seq(0,1,by=0.01))

##          0%          1%          2%          3%          4%          5% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##          6%          7%          8%          9%         10%         11% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         12%         13%         14%         15%         16%         17% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         18%         19%         20%         21%         22%         23% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         24%         25%         26%         27%         28%         29% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         30%         31%         32%         33%         34%         35% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         36%         37%         38%         39%         40%         41% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         42%         43%         44%         45%         46%         47% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         48%         49%         50%         51%         52%         53% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         54%         55%         56%         57%         58%         59% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         60%         61%         62%         63%         64%         65% 
##        0.00        0.00        0.00        0.00        0.00        0.00 
##         66%         67%         68%         69%         70%         71% 
##        0.00        0.00    12200.00    20000.00    31920.00    31920.00 
##         72%         73%         74%         75%         76%         77% 
##    33950.00    33950.00    33950.00    61884.25   127762.36   189655.46 
##         78%         79%         80%         81%         82%         83% 
##   238181.48   453527.86   524195.40   602165.00   803614.04   954853.36 
##         84%         85%         86%         87%         88%         89% 
##  1107856.68  1183850.15  1401638.64  1483828.00  1584879.52  1677979.52 
##         90%         91%         92%         93%         94%         95% 
##  1727225.00  1954311.91  2315173.44  2523524.00  3057532.00  3447819.00 
##         96%         97%         98%         99%        100% 
##  4709415.00  5000000.00  5132114.72  7882268.24 10050161.00

x <- subset(data$dh_val_pago_tarj_d, data$dh_val_pago_tarj_d <= 10560315.0)
hist(x)

dh_val_pago_tarj_d <- data$dh_val_pago_tarj_d

names(data)[113] # pc_productos_no_rotativos_entidad: Cantidad de productos no rotativos en el banco

## [1] "pc_productos_no_rotativos_entidad"

table(data$pc_productos_no_rotativos_entidad) 

## 
##   0   1 
## 901   1

hist(data$pc_productos_no_rotativos_entidad)

names(data)[114] # pc_saldo_no_rot_ent: Obligaciones no rotativos en el banco

## [1] "pc_saldo_no_rot_ent"

summary(data$pc_saldo_no_rot_ent)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##        0        0        0    40895        0 36887000

names(data)[115] # pc_vi_no_rotativos_entidad: Obligaciones no rotativos en el banco

## [1] "pc_vi_no_rotativos_entidad"

summary(data$pc_vi_no_rotativos_entidad)

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##     0.00     0.00     0.00    49.89     0.00 45000.00

names(data)[116] # dh_min_dia_entradas

## [1] "dh_min_dia_entradas"

names(data)[117] # pcons_hipotecario_vivienda: Es el producto un hipotecario

## [1] "pcons_hipotecario_vivienda"

table(data$pcons_hipotecario_vivienda)

## 
##   0   1 
## 879  23

pcons_hipotecario_vivienda <- data$pcons_hipotecario_vivienda

names(data)[118] # gsm_mejor_gestion_3m: Mejor gestión realizada en los últimos 3 meses

## [1] "gsm_mejor_gestion_3m"

table(data$gsm_mejor_gestion_3m)

## 
##   0  16 
## 901   1

names(data)[119] # dh_avg_dia_salidas: Dia promedio del mes en el que hace las salidas de dinero

## [1] "dh_avg_dia_salidas"

names(data)[120] # y_auto_cura: Variable respuesta: si el cliente se autocuro o no

## [1] "llave"

table(data$y_auto_cura)

## < table of extent 0 >

Y <- data$y_auto_cura

names(data)[122] # segmentoestructural: Segmento  estructural

## [1] "subsegmentoestructural"

table(data$segmentoestructural)

## 
## PYME 
##  902

names(data)[123] # subsegmentoestructural

## [1] "anhomes_ciclo"

table(data$subsegmentoestructural)

## 
## Pyme Pequena 
##          902

KEY <- data$llave
length(KEY)

## [1] 902

# Unificar variables

unificado <- as.data.frame(
  cbind(
    max_sem,
    desv_sem,
    prom_bim,
    max_mes_anterior,
    prom_mes_anterior,
    prom_sem,
    max_bim,
    prom_trim,
    pc_cant_moras_30_ult_12_meses,
    desv_trim,
    desv_bim,
    dh_cant_entradas,
    pc_transaccional,
    dh_val_entradas,
    pcons_tarjeta_de_credito,
    dh_avg_dia_retiros_d,
    dmi_max_egreso_diario,
    dh_cant_otros_d,
    dmi_max_ingreso_diario,
    dh_val_otros_d,
    pc_ingreso_final,
    dh_cant_pagos_d,
    dmi_ingreso_total_mes,
    dh_val_pagos_d,
    pc_gasto_familiar,
    pc_cuotas_pagadas,
    pc_ingreso_rutina_con_techo,
    dh_cant_salidas,
    dh_min_dia_pagos_d,
    pc_ingreso_por_rutina,
    dmi_egreso_total_mes,
    dh_val_salidas,
    cpc_avg_nro_cuota,
    cpc_sum_saldo,
    cpc_saldo_sobre_ing,
    dh_min_dia_pago_cred_d,
    cpc_saldo_tdc,
    dh_cant_tras_d,
    cpc_avg_saldo,
    dh_val_pago_tarj_d,
    pcons_hipotecario_vivienda,
    flag_mora60_ult12meses,
    flag_ultima_entrada_1quincena,
    flag_tuvo_mora60_ult3meses,
    flag_diaPago_1quincena,
    flag_es_cluster_6,
    flag_tuvo_mora90_ult12M,
    flax_maxDiaOtrosD_1quincena,
    flag_maxDiaPagos_d_1quincena,
    flag_encontrado_cifin
  )
)
dim(unificado)

## [1] 902  50

names(unificado)

##  [1] "max_sem"                       "desv_sem"                     
##  [3] "prom_bim"                      "max_mes_anterior"             
##  [5] "prom_mes_anterior"             "prom_sem"                     
##  [7] "max_bim"                       "prom_trim"                    
##  [9] "pc_cant_moras_30_ult_12_meses" "desv_trim"                    
## [11] "desv_bim"                      "dh_cant_entradas"             
## [13] "pc_transaccional"              "dh_val_entradas"              
## [15] "pcons_tarjeta_de_credito"      "dh_avg_dia_retiros_d"         
## [17] "dmi_max_egreso_diario"         "dh_cant_otros_d"              
## [19] "dmi_max_ingreso_diario"        "dh_val_otros_d"               
## [21] "pc_ingreso_final"              "dh_cant_pagos_d"              
## [23] "dmi_ingreso_total_mes"         "dh_val_pagos_d"               
## [25] "pc_gasto_familiar"             "pc_cuotas_pagadas"            
## [27] "pc_ingreso_rutina_con_techo"   "dh_cant_salidas"              
## [29] "dh_min_dia_pagos_d"            "pc_ingreso_por_rutina"        
## [31] "dmi_egreso_total_mes"          "dh_val_salidas"               
## [33] "cpc_avg_nro_cuota"             "cpc_sum_saldo"                
## [35] "cpc_saldo_sobre_ing"           "dh_min_dia_pago_cred_d"       
## [37] "cpc_saldo_tdc"                 "dh_cant_tras_d"               
## [39] "cpc_avg_saldo"                 "dh_val_pago_tarj_d"           
## [41] "pcons_hipotecario_vivienda"    "flag_mora60_ult12meses"       
## [43] "flag_ultima_entrada_1quincena" "flag_tuvo_mora60_ult3meses"   
## [45] "flag_diaPago_1quincena"        "flag_es_cluster_6"            
## [47] "flag_tuvo_mora90_ult12M"       "flax_maxDiaOtrosD_1quincena"  
## [49] "flag_maxDiaPagos_d_1quincena"  "flag_encontrado_cifin"

nrow(unificado)

## [1] 902

length(KEY)

## [1] 902

# Dejar solamente las variables que necesito
unificado <- unificado[, which(names(unificado) %in% c(
  'desv_sem',
  'max_mes_anterior',
  'desv_trim',
  'dh_cant_entradas',
  'pcons_tarjeta_de_credito',
  'dh_avg_dia_retiros_d',
  'dh_cant_otros_d',
  'dmi_max_ingreso_diario',
  'dh_val_otros_d',
  'dh_cant_pagos_d',
  'dh_val_pagos_d',
  'pc_cuotas_pagadas',
  'dh_min_dia_pagos_d',
  'pc_ingreso_por_rutina',
  'cpc_saldo_sobre_ing',
  'dh_min_dia_pago_cred_d',
  'cpc_saldo_tdc',
  'dh_cant_tras_d',
  'cpc_avg_saldo',
  'dh_val_pago_tarj_d',
  'flag_tuvo_mora60_ult3meses',
  'flag_diaPago_1quincena',
  'flag_es_cluster_6',
  'flag_tuvo_mora90_ult12M',
  'flax_maxDiaOtrosD_1quincena',
  'flag_encontrado_cifin'
))]
dim(unificado)

## [1] 902  26

names(unificado)

##  [1] "desv_sem"                    "max_mes_anterior"           
##  [3] "desv_trim"                   "dh_cant_entradas"           
##  [5] "pcons_tarjeta_de_credito"    "dh_avg_dia_retiros_d"       
##  [7] "dh_cant_otros_d"             "dmi_max_ingreso_diario"     
##  [9] "dh_val_otros_d"              "dh_cant_pagos_d"            
## [11] "dh_val_pagos_d"              "pc_cuotas_pagadas"          
## [13] "dh_min_dia_pagos_d"          "pc_ingreso_por_rutina"      
## [15] "cpc_saldo_sobre_ing"         "dh_min_dia_pago_cred_d"     
## [17] "cpc_saldo_tdc"               "dh_cant_tras_d"             
## [19] "cpc_avg_saldo"               "dh_val_pago_tarj_d"         
## [21] "flag_tuvo_mora60_ult3meses"  "flag_diaPago_1quincena"     
## [23] "flag_es_cluster_6"           "flag_tuvo_mora90_ult12M"    
## [25] "flax_maxDiaOtrosD_1quincena" "flag_encontrado_cifin"

# dejar solo variables independientes
KEY

##   [1]    1    3    4    6    7    8   10   11   13   14   16   18   19   23   24
##  [16]   25   28   30   32   33   39   40   42   43   46   48   50   51   55   56
##  [31]   57   58   59   60   61   63   65   66   67   70   75   78   87   88   89
##  [46]   93   94   95   96   97   99  100  101  103  104  105  106  107  108  109
##  [61]  111  114  116  119  120  121  122  124  125  126  127  129  130  131  132
##  [76]  133  134  136  138  141  142  143  144  145  147  148  149  152  153  154
##  [91]  155  157  160  163  164  165  169  170  171  177  182  184  185  187  189
## [106]  190  192  195  206  210  211  215  218  219  224  225  226  227  229  231
## [121]  234  235  237  238  241  242  243  245  246  249  250  251  252  254  259
## [136]  260  261  262  265  266  267  268  271  278  280  282  291  293  295  299
## [151]  302  304  305  306  312  313  314  319  321  322  324  326  329  331  334
## [166]  338  342  344  345  348  350  357  359  362  364  365  366  369  371  372
## [181]  373  374  378  383  384  385  386  387  390  391  395  398  400  404  407
## [196]  411  412  413  414  415  420  422  423  424  425  428  433  435  436  437
## [211]  438  440  444  450  452  456  461  462  465  467  468  476  479  482  483
## [226]  484  485  487  488  489  490  492  493  495  497  498  500  504  505  515
## [241]  516  517  530  533  534  537  539  540  541  543  546  547  549  551  552
## [256]  553  554  556  557  558  559  562  563  564  568  572  574  576  577  578
## [271]  581  583  585  586  590  592  593  601  606  607  609  613  614  615  616
## [286]  619  620  621  622  628  629  632  633  635  636  638  640  642  643  645
## [301]  649  650  653  654  657  661  662  664  668  669  671  673  675  676  677
## [316]  680  681  683  685  687  688  689  690  691  694  695  696  704  708  709
## [331]  715  729  730  732  733  734  738  739  746  749  750  757  761  764  765
## [346]  775  781  784  785  787  788  790  792  794  795  798  813  815  816  817
## [361]  818  819  820  821  822  824  826  827  830  831  836  837  838  840  841
## [376]  843  844  849  852  853  854  855  858  865  873  876  885  886  887  893
## [391]  894  901  902  906  907  912  913  914  915  920  921  924  925  926  927
## [406]  931  934  939  941  943  944  949  950  951  954  956  961  962  966  967
## [421]  968  970  973  974  975  977  980  984  986  988  991  993  994  995  998
## [436] 1000 1008 1009 1011 1012 1013 1014 1020 1021 1022 1024 1028 1030 1031 1033
## [451] 1035 1036 1044 1049 1052 1059 1061 1067 1068 1071 1073 1075 1076 1077 1080
## [466] 1082 1083 1085 1086 1088 1092 1093 1095 1096 1098 1099 1107 1108 1110 1111
## [481] 1112 1118 1120 1125 1127 1129 1132 1133 1134 1135 1137 1139 1140 1142 1144
## [496] 1145 1147 1150 1151 1154 1155 1156 1157 1158 1159 1168 1172 1176 1177 1179
## [511] 1182 1185 1186 1193 1195 1196 1199 1205 1206 1207 1210 1213 1215 1216 1218
## [526] 1219 1221 1223 1226 1227 1229 1235 1237 1240 1242 1243 1246 1247 1249 1250
## [541] 1254 1255 1256 1257 1261 1265 1266 1268 1269 1271 1272 1274 1275 1276 1281
## [556] 1292 1304 1313 1314 1315 1316 1317 1320 1321 1322 1323 1324 1328 1330 1331
## [571] 1332 1333 1334 1335 1339 1340 1342 1345 1347 1353 1355 1357 1361 1362 1369
## [586] 1371 1384 1389 1390 1391 1392 1393 1395 1398 1401 1406 1407 1408 1411 1412
## [601] 1416 1420 1424 1426 1430 1435 1436 1437 1438 1439 1440 1442 1443 1444 1447
## [616] 1449 1452 1454 1455 1457 1459 1462 1464 1465 1466 1469 1471 1474 1476 1477
## [631] 1479 1481 1484 1486 1488 1491 1493 1500 1501 1505 1508 1509 1511 1513 1514
## [646] 1519 1522 1523 1524 1527 1528 1529 1530 1533 1535 1542 1543 1544 1545 1548
## [661] 1550 1551 1554 1555 1556 1557 1559 1561 1563 1564 1567 1569 1572 1573 1574
## [676] 1576 1577 1578 1579 1580 1582 1583 1584 1587 1588 1589 1591 1592 1594 1597
## [691] 1599 1601 1602 1605 1606 1607 1609 1610 1611 1613 1615 1618 1619 1620 1622
## [706] 1625 1628 1630 1634 1635 1637 1638 1641 1643 1645 1646 1648 1649 1650 1656
## [721] 1660 1663 1664 1667 1669 1671 1673 1675 1679 1680 1683 1684 1687 1689 1690
## [736] 1693 1696 1699 1700 1704 1707 1708 1710 1711 1715 1716 1717 1718 1721 1722
## [751] 1725 1728 1729 1730 1731 1732 1733 1734 1735 1737 1738 1742 1744 1747 1749
## [766] 1753 1756 1758 1759 1761 1762 1765 1768 1769 1770 1771 1772 1777 1779 1780
## [781] 1781 1783 1784 1786 1787 1789 1792 1794 1795 1797 1802 1803 1806 1808 1809
## [796] 1810 1811 1812 1813 1820 1821 1822 1824 1825 1829 1831 1833 1834 1837 1840
## [811] 1847 1850 1852 1854 1864 1865 1866 1867 1871 1872 1873 1880 1883 1884 1885
## [826] 1889 1895 1896 1898 1899 1904 1905 1906 1907 1908 1909 1910 1911 1913 1916
## [841] 1918 1921 1922 1924 1925 1928 1929 1933 1935 1936 1938 1939 1941 1942 1943
## [856] 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1959 1964
## [871] 1967 1968 1969 1970 1974 1976 1977 1978 1984 1987 1988 1989 1990 1991 1992
## [886] 1993 1994 1997 1998 2000 2005 2007 2010 2012 2013 2014 2017 2019 2021 2025
## [901] 2027 2028

length(KEY)

## [1] 902

X <- unificado
X <- predict(preprocesado, X) # Aplicar Scaling
head(X)

##      desv_sem max_mes_anterior  desv_trim dh_cant_entradas
## 1 -0.25529797      -0.50752897 -1.1941103      -1.06028059
## 2  0.04710176       0.65178349  0.7213348      -1.06028059
## 3  0.01993882      -0.03016501  0.3804477      -1.06028059
## 4  0.65106628       1.94748564  0.8481227       0.07513835
## 5  0.47165449       0.71997834  0.7897396      -0.45337364
## 6  0.73660743      -0.50752897  0.8408677      -0.20198454
##   pcons_tarjeta_de_credito dh_avg_dia_retiros_d dh_cant_otros_d
## 1               -0.6752024           -1.2211367      -0.8645237
## 2               -0.6752024           -1.2211367      -0.8645237
## 3               -0.6752024           -1.2211367      -0.8645237
## 4               -0.6752024            0.5306734       0.8224426
## 5               -0.6752024            0.7530771      -0.8645237
## 6               -0.6752024            0.3631089      -0.1100894
##   dmi_max_ingreso_diario dh_val_otros_d dh_cant_pagos_d dh_val_pagos_d
## 1             -0.5939123     -0.4234230     -0.77779963     -0.5468244
## 2             -0.5939123     -0.4234230     -0.77779963     -0.5468244
## 3             -0.5939123     -0.4234230     -0.77779963     -0.5468244
## 4             -0.3281884     -0.3711720     -0.45552275      0.5283690
## 5             -0.5179912     -0.4234230     -0.13324587     -0.1850389
## 6             -0.2744742     -0.4150344      0.02789256      0.2742168
##   pc_cuotas_pagadas dh_min_dia_pagos_d pc_ingreso_por_rutina
## 1        -0.6194104         -0.6551503            -0.6333276
## 2        -0.6194104         -0.6551503            -0.6333276
## 3        -0.6194104         -0.6551503            -0.3666269
## 4        -0.6194104          0.5026302            -0.6248345
## 5        -0.6194104         -0.3243559            -0.6248345
## 6        -0.6194104         -0.4897531            -0.4254765
##   cpc_saldo_sobre_ing dh_min_dia_pago_cred_d cpc_saldo_tdc dh_cant_tras_d
## 1          -0.3699031             -0.6821349    -0.6288475     -0.4777662
## 2          -0.3699031             -0.6821349    -0.6288475     -0.4777662
## 3          -0.3699031             -0.6821349    -0.6288475     -0.4777662
## 4          -0.3699031              0.2916044    -0.6288475     -0.4777662
## 5          -0.3699031             -0.4039237    -0.6288475     -0.4777662
## 6          -0.3699031             -0.5430293    -0.6288475     -0.4777662
##   cpc_avg_saldo dh_val_pago_tarj_d flag_tuvo_mora60_ult3meses
## 1    -0.4360925         -0.3932686                 -0.3840753
## 2    -0.4360925         -0.3932686                 -0.3840753
## 3    -0.4360925         -0.3932686                 -0.3840753
## 4    -0.4360925         -0.3932686                 -0.3840753
## 5    -0.4360925         -0.3932686                 -0.3840753
## 6    -0.4360925         -0.3932686                 -0.3840753
##   flag_diaPago_1quincena flag_es_cluster_6 flag_tuvo_mora90_ult12M
## 1              0.6415816         -1.008661              -0.2625406
## 2              0.6415816         -1.008661              -0.2625406
## 3              0.6415816         -1.008661              -0.2625406
## 4              0.6415816         -1.008661              -0.2625406
## 5              0.6415816         -1.008661              -0.2625406
## 6              0.6415816         -1.008661              -0.2625406
##   flax_maxDiaOtrosD_1quincena flag_encontrado_cifin
## 1                    0.881491            -1.4521007
## 2                    0.881491            -1.4521007
## 3                    0.881491             0.6886098
## 4                   -1.134363             0.6886098
## 5                    0.881491             0.6886098
## 6                    0.881491             0.6886098

OOT <- cbind(KEY, X)
head(OOT)

##   KEY    desv_sem max_mes_anterior  desv_trim dh_cant_entradas
## 1   1 -0.25529797      -0.50752897 -1.1941103      -1.06028059
## 2   3  0.04710176       0.65178349  0.7213348      -1.06028059
## 3   4  0.01993882      -0.03016501  0.3804477      -1.06028059
## 4   6  0.65106628       1.94748564  0.8481227       0.07513835
## 5   7  0.47165449       0.71997834  0.7897396      -0.45337364
## 6   8  0.73660743      -0.50752897  0.8408677      -0.20198454
##   pcons_tarjeta_de_credito dh_avg_dia_retiros_d dh_cant_otros_d
## 1               -0.6752024           -1.2211367      -0.8645237
## 2               -0.6752024           -1.2211367      -0.8645237
## 3               -0.6752024           -1.2211367      -0.8645237
## 4               -0.6752024            0.5306734       0.8224426
## 5               -0.6752024            0.7530771      -0.8645237
## 6               -0.6752024            0.3631089      -0.1100894
##   dmi_max_ingreso_diario dh_val_otros_d dh_cant_pagos_d dh_val_pagos_d
## 1             -0.5939123     -0.4234230     -0.77779963     -0.5468244
## 2             -0.5939123     -0.4234230     -0.77779963     -0.5468244
## 3             -0.5939123     -0.4234230     -0.77779963     -0.5468244
## 4             -0.3281884     -0.3711720     -0.45552275      0.5283690
## 5             -0.5179912     -0.4234230     -0.13324587     -0.1850389
## 6             -0.2744742     -0.4150344      0.02789256      0.2742168
##   pc_cuotas_pagadas dh_min_dia_pagos_d pc_ingreso_por_rutina
## 1        -0.6194104         -0.6551503            -0.6333276
## 2        -0.6194104         -0.6551503            -0.6333276
## 3        -0.6194104         -0.6551503            -0.3666269
## 4        -0.6194104          0.5026302            -0.6248345
## 5        -0.6194104         -0.3243559            -0.6248345
## 6        -0.6194104         -0.4897531            -0.4254765
##   cpc_saldo_sobre_ing dh_min_dia_pago_cred_d cpc_saldo_tdc dh_cant_tras_d
## 1          -0.3699031             -0.6821349    -0.6288475     -0.4777662
## 2          -0.3699031             -0.6821349    -0.6288475     -0.4777662
## 3          -0.3699031             -0.6821349    -0.6288475     -0.4777662
## 4          -0.3699031              0.2916044    -0.6288475     -0.4777662
## 5          -0.3699031             -0.4039237    -0.6288475     -0.4777662
## 6          -0.3699031             -0.5430293    -0.6288475     -0.4777662
##   cpc_avg_saldo dh_val_pago_tarj_d flag_tuvo_mora60_ult3meses
## 1    -0.4360925         -0.3932686                 -0.3840753
## 2    -0.4360925         -0.3932686                 -0.3840753
## 3    -0.4360925         -0.3932686                 -0.3840753
## 4    -0.4360925         -0.3932686                 -0.3840753
## 5    -0.4360925         -0.3932686                 -0.3840753
## 6    -0.4360925         -0.3932686                 -0.3840753
##   flag_diaPago_1quincena flag_es_cluster_6 flag_tuvo_mora90_ult12M
## 1              0.6415816         -1.008661              -0.2625406
## 2              0.6415816         -1.008661              -0.2625406
## 3              0.6415816         -1.008661              -0.2625406
## 4              0.6415816         -1.008661              -0.2625406
## 5              0.6415816         -1.008661              -0.2625406
## 6              0.6415816         -1.008661              -0.2625406
##   flax_maxDiaOtrosD_1quincena flag_encontrado_cifin
## 1                    0.881491            -1.4521007
## 2                    0.881491            -1.4521007
## 3                    0.881491             0.6886098
## 4                   -1.134363             0.6886098
## 5                    0.881491             0.6886098
## 6                    0.881491             0.6886098

# Probabilidad dataset OOT
KEY <- OOT$KEY
probabilidad <- as.numeric(predict(RF2, OOT, type = "prob")[,2])
hist(probabilidad)

respuesta <- ifelse(probabilidad >= 0.50, 1, 0)
table(respuesta)

## respuesta
##   0   1 
## 379 523

llave <- KEY
imprimir <- as.data.frame(cbind(llave, probabilidad, respuesta))
head(imprimir)

##   llave probabilidad respuesta
## 1     1         0.42         0
## 2     3         0.46         0
## 3     4         0.26         0
## 4     6         0.25         0
## 5     7         0.31         0
## 6     8         0.40         0

table(imprimir$respuesta)

## 
##   0   1 
## 379 523

# write.csv(imprimir,"Base_prueba_evaluado.csv")