Script
Modelos Tesis de Grado 
Versión 1 - Balanceo de Datos y Reducción de Dimensionalidad
Hermilson Cruz Valencia- Michael Steven Morales - Hernán A. Tunjuelo Martínez
02/3/2023
APRENDIZAJE SUPERVISADO
1. Etapa 1: Entendimiento de los datos
Bases de Datos (Estructurada)
library(readxl)
DatosCopiaExp <- read_excel("D:/01_PUJ/00_Trabajo_de_grado/Data/DatosBD27052023.xlsx")
datos <- as.data.frame(DatosCopiaExp)
datos$Default=as.factor(datos$Default)
attach(datos)head(datos)Convertimos la variable Default en factor.
datos$Default <- factor(datos$Default)Tratamiento NA
colSums(is.na(datos))## Edad Estrato
## 0 0
## Personas a Cargo Tiemp_Vinc_Lab_Mes
## 0 0
## Ingreso_Basico Aux_Transporte
## 0 0
## Otros_Ingresos Base_70_Ingr_consti_Sal
## 0 0
## Ded_Nomina Tot_Gast_Fami
## 0 0
## Cuot_Men_Tdc_Cred_Rot Cuot_Mens_Otras_Oblig
## 0 0
## Sal_Disp Sal_Obli_Exter
## 0 0
## Sal_Obli_Ded_Nomina Sal_Créd_Viv_Veh
## 0 0
## Sal_TDC_Cred_Rot Punt_Acierta
## 0 0
## Tot_Cuot_Pag Porc_Proy_Liquid_Resp_Credito
## 0 0
## Porc_Ded_Nom Porc_Cap_Pago
## 0 0
## Porc_Gast_Fina End_Max_ Vec_Salario
## 0 0
## Mont_Credito Plazo_OPA
## 0 0
## CateAfil sexo2
## 0 0
## EstaCivil ConyLabora
## 0 0
## TipoVivienda NivelEduc
## 0 0
## TipoContrato CumpleTiempoVincul
## 0 0
## MoraObliVig Default
## 0 0str(datos)## 'data.frame': 15813 obs. of 36 variables:
## $ Edad : num 40 57 31 52 33 53 29 41 23 43 ...
## $ Estrato : num 3 2 3 2 2 1 3 3 3 1 ...
## $ Personas a Cargo : num 2 0 0 1 1 1 3 1 0 3 ...
## $ Tiemp_Vinc_Lab_Mes : num 23 57.5 26.3 143.6 11.4 ...
## $ Ingreso_Basico : num 737717 750000 1595984 2311221 737717 ...
## $ Aux_Transporte : num 83140 83140 0 0 83140 ...
## $ Otros_Ingresos : num 158200 0 0 0 0 ...
## $ Base_70_Ingr_consti_Sal : num 0 0 0 0 0 0 0 0 0 0 ...
## $ Ded_Nomina : num 59018 60000 129603 830024 120569 ...
## $ Tot_Gast_Fami : num 293717 249942 478795 693366 246257 ...
## $ Cuot_Men_Tdc_Cred_Rot : num 0 0 0 25254 0 ...
## $ Cuot_Mens_Otras_Oblig : num 0 0 0 0 0 181000 0 0 67000 60000 ...
## $ Sal_Disp : num 626322 523198 987586 762577 454031 ...
## $ Sal_Obli_Exter : num 0 0 0 5131006 0 ...
## $ Sal_Obli_Ded_Nomina : num 0 0 0 24638000 0 ...
## $ Sal_Créd_Viv_Veh : num 0 0 0 0 0 0 0 0 0 0 ...
## $ Sal_TDC_Cred_Rot : num 0 0 0 467000 0 0 0 0 0 0 ...
## $ Punt_Acierta : num 683 803 838 665 794 694 424 705 727 556 ...
## $ Tot_Cuot_Pag : num 96516 72842 107736 116714 100158 ...
## $ Porc_Proy_Liquid_Resp_Credito: num 1.121 3.309 0.954 0 0.671 ...
## $ Porc_Ded_Nom : num 0.211 0.177 0.149 0.41 0.299 ...
## $ Porc_Cap_Pago : num 0.459 0.459 0.449 0.721 0.569 ...
## $ Porc_Gast_Fina : num 0.1049 0.0942 0.0735 0.0958 0.143 ...
## $ End_Max_ Vec_Salario : num 1.08 0.96 1.88 14.29 1.34 ...
## $ Mont_Credito : num 1060000 800000 3000000 3250000 1100000 ...
## $ Plazo_OPA : num 12 12 36 36 12 12 12 15 18 18 ...
## $ CateAfil : num 1 1 3 3 1 2 2 1 2 2 ...
## $ sexo2 : num 1 1 1 0 0 0 1 1 1 1 ...
## $ EstaCivil : num 3 2 3 1 1 1 3 1 1 3 ...
## $ ConyLabora : num 1 1 1 1 1 1 2 1 1 2 ...
## $ TipoVivienda : num 2 3 1 3 1 2 3 1 2 2 ...
## $ NivelEduc : num 2 2 3 5 2 1 2 2 3 2 ...
## $ TipoContrato : num 7 7 7 7 7 7 2 7 7 7 ...
## $ CumpleTiempoVincul : num 2 2 2 2 2 2 2 2 2 2 ...
## $ MoraObliVig : num 6 15 15 15 15 15 10 15 15 15 ...
## $ Default : Factor w/ 2 levels "0","1": 1 1 1 1 1 1 1 2 1 2 ...library(skimr)
skimr::skim(datos)| Name | datos |
| Number of rows | 15813 |
| Number of columns | 36 |
| _______________________ | |
| Column type frequency: | |
| factor | 1 |
| numeric | 35 |
| ________________________ | |
| Group variables | None |
Variable type: factor
| skim_variable | n_missing | complete_rate | ordered | n_unique | top_counts |
|---|---|---|---|---|---|
| Default | 0 | 1 | FALSE | 2 | 0: 15190, 1: 623 |
Variable type: numeric
| skim_variable | n_missing | complete_rate | mean | sd | p0 | p25 | p50 | p75 | p100 | hist |
|---|---|---|---|---|---|---|---|---|---|---|
| Edad | 0 | 1 | 35.72 | 1.060000e+01 | 18.00 | 27.00 | 34.00 | 43.00 | 6.900000e+01 | ▇▇▅▃▁ |
| Estrato | 0 | 1 | 2.29 | 8.800000e-01 | 1.00 | 2.00 | 2.00 | 3.00 | 6.000000e+00 | ▇▅▁▁▁ |
| Personas a Cargo | 0 | 1 | 1.18 | 1.130000e+00 | 0.00 | 0.00 | 1.00 | 2.00 | 1.000000e+01 | ▇▁▁▁▁ |
| Tiemp_Vinc_Lab_Mes | 0 | 1 | 60.74 | 6.933000e+01 | 0.03 | 16.97 | 34.80 | 73.87 | 4.771700e+02 | ▇▁▁▁▁ |
| Ingreso_Basico | 0 | 1 | 1294322.86 | 9.401741e+05 | 737717.00 | 828116.00 | 953614.00 | 1418970.00 | 2.544700e+07 | ▇▁▁▁▁ |
| Aux_Transporte | 0 | 1 | 71771.77 | 4.112031e+04 | 0.00 | 83140.00 | 88211.00 | 97032.00 | 1.028540e+05 | ▂▁▁▁▇ |
| Otros_Ingresos | 0 | 1 | 47443.51 | 3.970631e+05 | 0.00 | 0.00 | 0.00 | 0.00 | 1.933709e+07 | ▇▁▁▁▁ |
| Base_70_Ingr_consti_Sal | 0 | 1 | 96509.35 | 3.020664e+05 | 0.00 | 0.00 | 0.00 | 0.00 | 7.059080e+06 | ▇▁▁▁▁ |
| Ded_Nomina | 0 | 1 | 202413.09 | 3.140426e+05 | 0.00 | 76827.00 | 121474.16 | 221600.00 | 1.148501e+07 | ▇▁▁▁▁ |
| Tot_Gast_Fami | 0 | 1 | 382546.43 | 2.831822e+05 | 110657.55 | 260835.90 | 315983.40 | 428592.00 | 9.028440e+06 | ▇▁▁▁▁ |
| Cuot_Men_Tdc_Cred_Rot | 0 | 1 | 15317.63 | 8.260993e+04 | 0.00 | 0.00 | 0.00 | 0.00 | 3.892214e+06 | ▇▁▁▁▁ |
| Cuot_Mens_Otras_Oblig | 0 | 1 | 143032.63 | 3.406665e+05 | 0.00 | 0.00 | 63000.00 | 181000.00 | 1.338300e+07 | ▇▁▁▁▁ |
| Sal_Disp | 0 | 1 | 737800.60 | 4.646815e+05 | -5283079.07 | 507679.10 | 628977.78 | 845198.00 | 1.697228e+07 | ▁▇▁▁▁ |
| Sal_Obli_Exter | 0 | 1 | 2156011.86 | 7.630374e+06 | 0.00 | 0.00 | 420000.00 | 1855000.00 | 3.234400e+08 | ▇▁▁▁▁ |
| Sal_Obli_Ded_Nomina | 0 | 1 | 501701.46 | 3.761631e+06 | 0.00 | 0.00 | 0.00 | 0.00 | 1.194650e+08 | ▇▁▁▁▁ |
| Sal_Créd_Viv_Veh | 0 | 1 | 12059185.81 | 1.298581e+09 | 0.00 | 0.00 | 0.00 | 0.00 | 1.632890e+11 | ▇▁▁▁▁ |
| Sal_TDC_Cred_Rot | 0 | 1 | 234124.38 | 1.280756e+06 | 0.00 | 0.00 | 0.00 | 0.00 | 5.733900e+07 | ▇▁▁▁▁ |
| Punt_Acierta | 0 | 1 | 683.64 | 1.862700e+02 | 0.00 | 667.00 | 727.00 | 781.00 | 9.380000e+02 | ▁▁▁▇▆ |
| Tot_Cuot_Pag | 0 | 1 | 174075.56 | 1.176718e+05 | 9663.16 | 108106.23 | 136874.08 | 190717.41 | 1.637944e+06 | ▇▁▁▁▁ |
| Porc_Proy_Liquid_Resp_Credito | 0 | 1 | 1.25 | 2.000000e+00 | 0.00 | 0.50 | 0.73 | 1.30 | 5.725000e+01 | ▇▁▁▁▁ |
| Porc_Ded_Nom | 0 | 1 | 0.27 | 9.000000e-02 | 0.06 | 0.20 | 0.25 | 0.33 | 1.160000e+00 | ▇▅▁▁▁ |
| Porc_Cap_Pago | 0 | 1 | 0.60 | 1.400000e-01 | 0.18 | 0.50 | 0.59 | 0.70 | 2.610000e+00 | ▇▃▁▁▁ |
| Porc_Gast_Fina | 0 | 1 | 0.23 | 1.200000e-01 | 0.01 | 0.14 | 0.21 | 0.31 | 2.430000e+00 | ▇▁▁▁▁ |
| End_Max_ Vec_Salario | 0 | 1 | 4.41 | 3.160000e+00 | 0.21 | 2.14 | 3.57 | 5.90 | 3.981000e+01 | ▇▁▁▁▁ |
| Mont_Credito | 0 | 1 | 4560092.87 | 4.782340e+06 | 340000.00 | 1700000.00 | 3053000.00 | 5346000.00 | 5.382754e+07 | ▇▁▁▁▁ |
| Plazo_OPA | 0 | 1 | 29.73 | 1.659000e+01 | 3.00 | 18.00 | 24.00 | 36.00 | 9.600000e+01 | ▇▇▂▃▁ |
| CateAfil | 0 | 1 | 1.98 | 7.400000e-01 | 1.00 | 2.00 | 2.00 | 2.00 | 4.000000e+00 | ▃▇▁▂▁ |
| sexo2 | 0 | 1 | 0.58 | 4.900000e-01 | 0.00 | 0.00 | 1.00 | 1.00 | 1.000000e+00 | ▆▁▁▁▇ |
| EstaCivil | 0 | 1 | 1.84 | 9.700000e-01 | 1.00 | 1.00 | 1.00 | 3.00 | 5.000000e+00 | ▇▃▃▁▁ |
| ConyLabora | 0 | 1 | 1.26 | 4.400000e-01 | 1.00 | 1.00 | 1.00 | 2.00 | 2.000000e+00 | ▇▁▁▁▃ |
| TipoVivienda | 0 | 1 | 2.11 | 6.400000e-01 | 1.00 | 2.00 | 2.00 | 3.00 | 3.000000e+00 | ▂▁▇▁▃ |
| NivelEduc | 0 | 1 | 2.86 | 1.310000e+00 | 1.00 | 2.00 | 2.00 | 4.00 | 6.000000e+00 | ▇▃▂▂▁ |
| TipoContrato | 0 | 1 | 5.32 | 2.450000e+00 | 1.00 | 2.00 | 7.00 | 7.00 | 7.000000e+00 | ▃▁▁▁▇ |
| CumpleTiempoVincul | 0 | 1 | 1.94 | 2.300000e-01 | 1.00 | 2.00 | 2.00 | 2.00 | 2.000000e+00 | ▁▁▁▁▇ |
| MoraObliVig | 0 | 1 | 14.23 | 2.360000e+00 | 1.00 | 15.00 | 15.00 | 15.00 | 1.500000e+01 | ▁▁▁▁▇ |
table(datos$Default)##
## 0 1
## 15190 623prop.table(table(datos$Default))##
## 0 1
## 0.96060204 0.039397962. Etapa 2: Preparación de los datos
La variable Default se cambia a factor
2.1. Modelo Logit sin reducción de dimensionalidad ni balanceo de datos
datos$Default=as.factor(datos$Default)
mod=glm(Default~.,data=datos,family = "binomial")
mod##
## Call: glm(formula = Default ~ ., family = "binomial", data = datos)
##
## Coefficients:
## (Intercept) Edad
## 2.322e+00 -1.520e-02
## Estrato `Personas a Cargo`
## -3.197e-02 -1.494e-02
## Tiemp_Vinc_Lab_Mes Ingreso_Basico
## 9.795e-04 6.057e-01
## Aux_Transporte Otros_Ingresos
## 6.057e-01 6.057e-01
## Base_70_Ingr_consti_Sal Ded_Nomina
## 4.240e-01 -6.057e-01
## Tot_Gast_Fami Cuot_Men_Tdc_Cred_Rot
## -6.057e-01 -6.058e-01
## Cuot_Mens_Otras_Oblig Sal_Disp
## -6.057e-01 -6.057e-01
## Sal_Obli_Exter Sal_Obli_Ded_Nomina
## 7.789e-09 -2.180e-08
## Sal_Créd_Viv_Veh Sal_TDC_Cred_Rot
## -5.111e-09 5.615e-07
## Punt_Acierta Tot_Cuot_Pag
## -4.705e-04 -7.474e-06
## Porc_Proy_Liquid_Resp_Credito Porc_Ded_Nom
## -4.830e-01 3.313e+00
## Porc_Cap_Pago Porc_Gast_Fina
## -5.469e+00 6.156e+00
## `End_Max_ Vec_Salario` Mont_Credito
## -4.696e-02 9.447e-08
## Plazo_OPA CateAfil
## -4.304e-03 -3.283e-01
## sexo2 EstaCivil
## 3.032e-01 2.909e-04
## ConyLabora TipoVivienda
## -8.615e-01 -6.178e-02
## NivelEduc TipoContrato
## -1.263e-01 -5.824e-02
## CumpleTiempoVincul MoraObliVig
## 1.970e-01 -3.150e-02
##
## Degrees of Freedom: 15812 Total (i.e. Null); 15777 Residual
## Null Deviance: 5251
## Residual Deviance: 4964 AIC: 5036summary(mod)##
## Call:
## glm(formula = Default ~ ., family = "binomial", data = datos)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.1716 -0.3235 -0.2598 -0.1931 3.7527
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 2.322e+00 1.013e+00 2.293 0.021871 *
## Edad -1.520e-02 5.013e-03 -3.033 0.002423 **
## Estrato -3.197e-02 5.694e-02 -0.561 0.574556
## `Personas a Cargo` -1.494e-02 4.105e-02 -0.364 0.715945
## Tiemp_Vinc_Lab_Mes 9.795e-04 1.763e-03 0.556 0.578506
## Ingreso_Basico 6.057e-01 8.035e-01 0.754 0.450905
## Aux_Transporte 6.057e-01 8.035e-01 0.754 0.450908
## Otros_Ingresos 6.057e-01 8.035e-01 0.754 0.450906
## Base_70_Ingr_consti_Sal 4.240e-01 5.624e-01 0.754 0.450906
## Ded_Nomina -6.057e-01 8.035e-01 -0.754 0.450906
## Tot_Gast_Fami -6.057e-01 8.035e-01 -0.754 0.450906
## Cuot_Men_Tdc_Cred_Rot -6.058e-01 8.035e-01 -0.754 0.450899
## Cuot_Mens_Otras_Oblig -6.057e-01 8.035e-01 -0.754 0.450906
## Sal_Disp -6.057e-01 8.035e-01 -0.754 0.450906
## Sal_Obli_Exter 7.789e-09 1.416e-08 0.550 0.582161
## Sal_Obli_Ded_Nomina -2.180e-08 2.752e-08 -0.792 0.428158
## Sal_Créd_Viv_Veh -5.111e-09 6.326e-09 -0.808 0.419144
## Sal_TDC_Cred_Rot 5.615e-07 3.872e-07 1.450 0.147036
## Punt_Acierta -4.705e-04 2.005e-04 -2.347 0.018947 *
## Tot_Cuot_Pag -7.474e-06 1.972e-06 -3.790 0.000151 ***
## Porc_Proy_Liquid_Resp_Credito -4.830e-01 1.075e-01 -4.494 6.99e-06 ***
## Porc_Ded_Nom 3.313e+00 1.537e+00 2.155 0.031167 *
## Porc_Cap_Pago -5.469e+00 1.938e+00 -2.822 0.004776 **
## Porc_Gast_Fina 6.156e+00 1.937e+00 3.177 0.001487 **
## `End_Max_ Vec_Salario` -4.696e-02 3.642e-02 -1.290 0.197204
## Mont_Credito 9.448e-08 4.325e-08 2.185 0.028917 *
## Plazo_OPA -4.304e-03 6.321e-03 -0.681 0.495943
## CateAfil -3.283e-01 1.043e-01 -3.149 0.001640 **
## sexo2 3.032e-01 9.266e-02 3.273 0.001065 **
## EstaCivil 2.909e-04 5.311e-02 0.005 0.995630
## ConyLabora -8.615e-01 3.357e-01 -2.566 0.010287 *
## TipoVivienda -6.178e-02 6.868e-02 -0.900 0.368369
## NivelEduc -1.263e-01 4.752e-02 -2.658 0.007868 **
## TipoContrato -5.824e-02 1.720e-02 -3.385 0.000711 ***
## CumpleTiempoVincul 1.970e-01 1.655e-01 1.191 0.233768
## MoraObliVig -3.150e-02 1.699e-02 -1.853 0.063828 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5250.7 on 15812 degrees of freedom
## Residual deviance: 4964.3 on 15777 degrees of freedom
## AIC: 5036.3
##
## Number of Fisher Scoring iterations: 17Se obtiene unas pocas variables con significancia.
2.2. Usando Setp para reduccion de dimensionalidad
La reducción de dimensionalidad se encarga de identificar y eliminar las características que disminuyen el rendimiento del modelo de aprendizaje automático.
mod_2=step(mod)## Start: AIC=5036.27
## Default ~ Edad + Estrato + `Personas a Cargo` + Tiemp_Vinc_Lab_Mes +
## Ingreso_Basico + Aux_Transporte + Otros_Ingresos + Base_70_Ingr_consti_Sal +
## Ded_Nomina + Tot_Gast_Fami + Cuot_Men_Tdc_Cred_Rot + Cuot_Mens_Otras_Oblig +
## Sal_Disp + Sal_Obli_Exter + Sal_Obli_Ded_Nomina + Sal_Créd_Viv_Veh +
## Sal_TDC_Cred_Rot + Punt_Acierta + Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito +
## Porc_Ded_Nom + Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + Plazo_OPA + CateAfil + sexo2 + EstaCivil +
## ConyLabora + TipoVivienda + NivelEduc + TipoContrato + CumpleTiempoVincul +
## MoraObliVig
##
## Df Deviance AIC
## - EstaCivil 1 4964.3 5034.3
## - `Personas a Cargo` 1 4964.4 5034.4
## - Sal_Obli_Exter 1 4964.6 5034.6
## - Tiemp_Vinc_Lab_Mes 1 4964.6 5034.6
## - Estrato 1 4964.6 5034.6
## - Plazo_OPA 1 4964.7 5034.7
## - Sal_Obli_Ded_Nomina 1 4965.0 5035.0
## - Sal_Créd_Viv_Veh 1 4965.1 5035.1
## - TipoVivienda 1 4965.1 5035.1
## - Aux_Transporte 1 4965.2 5035.2
## - Base_70_Ingr_consti_Sal 1 4965.2 5035.2
## - Cuot_Mens_Otras_Oblig 1 4965.3 5035.3
## - Otros_Ingresos 1 4965.3 5035.3
## - Sal_Disp 1 4965.3 5035.3
## - Tot_Gast_Fami 1 4965.3 5035.3
## - Ded_Nomina 1 4965.3 5035.3
## - Ingreso_Basico 1 4965.3 5035.3
## - Cuot_Men_Tdc_Cred_Rot 1 4965.5 5035.5
## - CumpleTiempoVincul 1 4965.7 5035.7
## - `End_Max_ Vec_Salario` 1 4966.0 5036.0
## - Sal_TDC_Cred_Rot 1 4966.2 5036.2
## <none> 4964.3 5036.3
## - MoraObliVig 1 4967.5 5037.5
## - Mont_Credito 1 4968.8 5038.8
## - Porc_Ded_Nom 1 4968.8 5038.8
## - Punt_Acierta 1 4969.5 5039.5
## - ConyLabora 1 4970.9 5040.9
## - NivelEduc 1 4971.5 5041.5
## - Porc_Cap_Pago 1 4972.2 5042.2
## - Edad 1 4973.8 5043.8
## - CateAfil 1 4974.2 5044.2
## - Porc_Gast_Fina 1 4974.2 5044.2
## - sexo2 1 4975.2 5045.2
## - TipoContrato 1 4975.5 5045.5
## - Tot_Cuot_Pag 1 4979.1 5049.1
## - Porc_Proy_Liquid_Resp_Credito 1 4990.9 5060.9
##
## Step: AIC=5034.27
## Default ~ Edad + Estrato + `Personas a Cargo` + Tiemp_Vinc_Lab_Mes +
## Ingreso_Basico + Aux_Transporte + Otros_Ingresos + Base_70_Ingr_consti_Sal +
## Ded_Nomina + Tot_Gast_Fami + Cuot_Men_Tdc_Cred_Rot + Cuot_Mens_Otras_Oblig +
## Sal_Disp + Sal_Obli_Exter + Sal_Obli_Ded_Nomina + Sal_Créd_Viv_Veh +
## Sal_TDC_Cred_Rot + Punt_Acierta + Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito +
## Porc_Ded_Nom + Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + Plazo_OPA + CateAfil + sexo2 + ConyLabora +
## TipoVivienda + NivelEduc + TipoContrato + CumpleTiempoVincul +
## MoraObliVig
##
## Df Deviance AIC
## - `Personas a Cargo` 1 4964.4 5032.4
## - Sal_Obli_Exter 1 4964.6 5032.6
## - Tiemp_Vinc_Lab_Mes 1 4964.6 5032.6
## - Estrato 1 4964.6 5032.6
## - Plazo_OPA 1 4964.7 5032.7
## - Sal_Obli_Ded_Nomina 1 4965.0 5033.0
## - Sal_Créd_Viv_Veh 1 4965.1 5033.1
## - TipoVivienda 1 4965.1 5033.1
## - Aux_Transporte 1 4965.2 5033.2
## - Base_70_Ingr_consti_Sal 1 4965.2 5033.2
## - Cuot_Mens_Otras_Oblig 1 4965.3 5033.3
## - Otros_Ingresos 1 4965.3 5033.3
## - Sal_Disp 1 4965.3 5033.3
## - Tot_Gast_Fami 1 4965.3 5033.3
## - Ded_Nomina 1 4965.3 5033.3
## - Ingreso_Basico 1 4965.3 5033.3
## - Cuot_Men_Tdc_Cred_Rot 1 4965.5 5033.5
## - CumpleTiempoVincul 1 4965.7 5033.7
## - `End_Max_ Vec_Salario` 1 4966.0 5034.0
## - Sal_TDC_Cred_Rot 1 4966.2 5034.2
## <none> 4964.3 5034.3
## - MoraObliVig 1 4967.5 5035.5
## - Mont_Credito 1 4968.8 5036.8
## - Porc_Ded_Nom 1 4968.8 5036.8
## - Punt_Acierta 1 4969.5 5037.5
## - ConyLabora 1 4971.2 5039.2
## - NivelEduc 1 4971.5 5039.5
## - Porc_Cap_Pago 1 4972.2 5040.2
## - Edad 1 4974.0 5042.0
## - CateAfil 1 4974.2 5042.2
## - Porc_Gast_Fina 1 4974.2 5042.2
## - sexo2 1 4975.2 5043.2
## - TipoContrato 1 4975.5 5043.5
## - Tot_Cuot_Pag 1 4979.1 5047.1
## - Porc_Proy_Liquid_Resp_Credito 1 4990.9 5058.9
##
## Step: AIC=5032.41
## Default ~ Edad + Estrato + Tiemp_Vinc_Lab_Mes + Ingreso_Basico +
## Aux_Transporte + Otros_Ingresos + Base_70_Ingr_consti_Sal +
## Ded_Nomina + Tot_Gast_Fami + Cuot_Men_Tdc_Cred_Rot + Cuot_Mens_Otras_Oblig +
## Sal_Disp + Sal_Obli_Exter + Sal_Obli_Ded_Nomina + Sal_Créd_Viv_Veh +
## Sal_TDC_Cred_Rot + Punt_Acierta + Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito +
## Porc_Ded_Nom + Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + Plazo_OPA + CateAfil + sexo2 + ConyLabora +
## TipoVivienda + NivelEduc + TipoContrato + CumpleTiempoVincul +
## MoraObliVig
##
## Df Deviance AIC
## - Estrato 1 4964.7 5030.7
## - Sal_Obli_Exter 1 4964.7 5030.7
## - Tiemp_Vinc_Lab_Mes 1 4964.7 5030.7
## - Plazo_OPA 1 4964.9 5030.9
## - Sal_Obli_Ded_Nomina 1 4965.1 5031.1
## - TipoVivienda 1 4965.2 5031.2
## - Sal_Créd_Viv_Veh 1 4965.2 5031.2
## - Aux_Transporte 1 4965.3 5031.3
## - Base_70_Ingr_consti_Sal 1 4965.4 5031.4
## - Cuot_Mens_Otras_Oblig 1 4965.4 5031.4
## - Otros_Ingresos 1 4965.4 5031.4
## - Sal_Disp 1 4965.4 5031.4
## - Tot_Gast_Fami 1 4965.4 5031.4
## - Ded_Nomina 1 4965.4 5031.4
## - Ingreso_Basico 1 4965.4 5031.4
## - Cuot_Men_Tdc_Cred_Rot 1 4965.6 5031.6
## - CumpleTiempoVincul 1 4965.9 5031.9
## - `End_Max_ Vec_Salario` 1 4966.1 5032.1
## - Sal_TDC_Cred_Rot 1 4966.4 5032.4
## <none> 4964.4 5032.4
## - MoraObliVig 1 4967.6 5033.6
## - Mont_Credito 1 4968.9 5034.9
## - Porc_Ded_Nom 1 4969.0 5035.0
## - Punt_Acierta 1 4969.7 5035.7
## - ConyLabora 1 4971.4 5037.4
## - NivelEduc 1 4971.5 5037.5
## - Porc_Cap_Pago 1 4972.4 5038.4
## - CateAfil 1 4974.3 5040.3
## - Porc_Gast_Fina 1 4974.4 5040.4
## - Edad 1 4975.2 5041.2
## - sexo2 1 4975.2 5041.2
## - TipoContrato 1 4975.6 5041.6
## - Tot_Cuot_Pag 1 4979.3 5045.3
## - Porc_Proy_Liquid_Resp_Credito 1 4991.2 5057.2
##
## Step: AIC=5030.67
## Default ~ Edad + Tiemp_Vinc_Lab_Mes + Ingreso_Basico + Aux_Transporte +
## Otros_Ingresos + Base_70_Ingr_consti_Sal + Ded_Nomina + Tot_Gast_Fami +
## Cuot_Men_Tdc_Cred_Rot + Cuot_Mens_Otras_Oblig + Sal_Disp +
## Sal_Obli_Exter + Sal_Obli_Ded_Nomina + Sal_Créd_Viv_Veh +
## Sal_TDC_Cred_Rot + Punt_Acierta + Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito +
## Porc_Ded_Nom + Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + Plazo_OPA + CateAfil + sexo2 + ConyLabora +
## TipoVivienda + NivelEduc + TipoContrato + CumpleTiempoVincul +
## MoraObliVig
##
## Df Deviance AIC
## - Sal_Obli_Exter 1 4964.9 5028.9
## - Tiemp_Vinc_Lab_Mes 1 4965.0 5029.0
## - Plazo_OPA 1 4965.1 5029.1
## - Sal_Obli_Ded_Nomina 1 4965.4 5029.4
## - TipoVivienda 1 4965.4 5029.4
## - Sal_Créd_Viv_Veh 1 4965.5 5029.5
## - Aux_Transporte 1 4965.6 5029.6
## - Base_70_Ingr_consti_Sal 1 4965.6 5029.6
## - Cuot_Mens_Otras_Oblig 1 4965.7 5029.7
## - Otros_Ingresos 1 4965.7 5029.7
## - Tot_Gast_Fami 1 4965.7 5029.7
## - Sal_Disp 1 4965.7 5029.7
## - Ded_Nomina 1 4965.7 5029.7
## - Ingreso_Basico 1 4965.7 5029.7
## - Cuot_Men_Tdc_Cred_Rot 1 4965.9 5029.9
## - CumpleTiempoVincul 1 4966.1 5030.1
## - `End_Max_ Vec_Salario` 1 4966.3 5030.3
## - Sal_TDC_Cred_Rot 1 4966.6 5030.6
## <none> 4964.7 5030.7
## - MoraObliVig 1 4967.8 5031.8
## - Mont_Credito 1 4969.1 5033.1
## - Porc_Ded_Nom 1 4969.2 5033.2
## - Punt_Acierta 1 4970.0 5034.0
## - ConyLabora 1 4971.6 5035.6
## - Porc_Cap_Pago 1 4972.5 5036.5
## - NivelEduc 1 4972.9 5036.9
## - Porc_Gast_Fina 1 4974.5 5038.5
## - CateAfil 1 4974.8 5038.8
## - sexo2 1 4975.6 5039.6
## - Edad 1 4975.9 5039.9
## - TipoContrato 1 4975.9 5039.9
## - Tot_Cuot_Pag 1 4979.5 5043.5
## - Porc_Proy_Liquid_Resp_Credito 1 4991.5 5055.5
##
## Step: AIC=5028.95
## Default ~ Edad + Tiemp_Vinc_Lab_Mes + Ingreso_Basico + Aux_Transporte +
## Otros_Ingresos + Base_70_Ingr_consti_Sal + Ded_Nomina + Tot_Gast_Fami +
## Cuot_Men_Tdc_Cred_Rot + Cuot_Mens_Otras_Oblig + Sal_Disp +
## Sal_Obli_Ded_Nomina + Sal_Créd_Viv_Veh + Sal_TDC_Cred_Rot +
## Punt_Acierta + Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito +
## Porc_Ded_Nom + Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + Plazo_OPA + CateAfil + sexo2 + ConyLabora +
## TipoVivienda + NivelEduc + TipoContrato + CumpleTiempoVincul +
## MoraObliVig
##
## Df Deviance AIC
## - Tiemp_Vinc_Lab_Mes 1 4965.2 5027.2
## - Plazo_OPA 1 4965.6 5027.6
## - TipoVivienda 1 4965.7 5027.7
## - Aux_Transporte 1 4965.9 5027.9
## - Base_70_Ingr_consti_Sal 1 4965.9 5027.9
## - Cuot_Mens_Otras_Oblig 1 4965.9 5027.9
## - Tot_Gast_Fami 1 4965.9 5027.9
## - Otros_Ingresos 1 4965.9 5027.9
## - Sal_Disp 1 4965.9 5027.9
## - Ded_Nomina 1 4965.9 5027.9
## - Ingreso_Basico 1 4965.9 5027.9
## - Sal_Obli_Ded_Nomina 1 4966.1 5028.1
## - Cuot_Men_Tdc_Cred_Rot 1 4966.1 5028.1
## - Sal_Créd_Viv_Veh 1 4966.3 5028.3
## - `End_Max_ Vec_Salario` 1 4966.4 5028.4
## - CumpleTiempoVincul 1 4966.4 5028.4
## - Sal_TDC_Cred_Rot 1 4966.8 5028.8
## <none> 4964.9 5028.9
## - MoraObliVig 1 4968.1 5030.1
## - Porc_Ded_Nom 1 4969.3 5031.3
## - Mont_Credito 1 4969.3 5031.3
## - Punt_Acierta 1 4970.3 5032.3
## - ConyLabora 1 4971.7 5033.7
## - Porc_Cap_Pago 1 4972.8 5034.8
## - NivelEduc 1 4973.2 5035.2
## - Porc_Gast_Fina 1 4974.5 5036.5
## - CateAfil 1 4975.1 5037.1
## - sexo2 1 4975.9 5037.9
## - TipoContrato 1 4976.2 5038.2
## - Edad 1 4976.2 5038.2
## - Tot_Cuot_Pag 1 4979.5 5041.5
## - Porc_Proy_Liquid_Resp_Credito 1 4991.5 5053.5
##
## Step: AIC=5027.22
## Default ~ Edad + Ingreso_Basico + Aux_Transporte + Otros_Ingresos +
## Base_70_Ingr_consti_Sal + Ded_Nomina + Tot_Gast_Fami + Cuot_Men_Tdc_Cred_Rot +
## Cuot_Mens_Otras_Oblig + Sal_Disp + Sal_Obli_Ded_Nomina +
## Sal_Créd_Viv_Veh + Sal_TDC_Cred_Rot + Punt_Acierta + Tot_Cuot_Pag +
## Porc_Proy_Liquid_Resp_Credito + Porc_Ded_Nom + Porc_Cap_Pago +
## Porc_Gast_Fina + `End_Max_ Vec_Salario` + Mont_Credito +
## Plazo_OPA + CateAfil + sexo2 + ConyLabora + TipoVivienda +
## NivelEduc + TipoContrato + CumpleTiempoVincul + MoraObliVig
##
## Df Deviance AIC
## - Plazo_OPA 1 4965.7 5025.7
## - TipoVivienda 1 4965.9 5025.9
## - Sal_Obli_Ded_Nomina 1 4966.1 5026.1
## - Aux_Transporte 1 4966.2 5026.2
## - Base_70_Ingr_consti_Sal 1 4966.2 5026.2
## - Cuot_Mens_Otras_Oblig 1 4966.2 5026.2
## - Tot_Gast_Fami 1 4966.2 5026.2
## - Otros_Ingresos 1 4966.2 5026.2
## - Sal_Disp 1 4966.2 5026.2
## - Ded_Nomina 1 4966.2 5026.2
## - Ingreso_Basico 1 4966.2 5026.2
## - Cuot_Men_Tdc_Cred_Rot 1 4966.4 5026.4
## - Sal_Créd_Viv_Veh 1 4966.5 5026.5
## - `End_Max_ Vec_Salario` 1 4966.5 5026.5
## - CumpleTiempoVincul 1 4966.6 5026.6
## - Sal_TDC_Cred_Rot 1 4967.1 5027.1
## <none> 4965.2 5027.2
## - MoraObliVig 1 4968.4 5028.4
## - Mont_Credito 1 4969.7 5029.7
## - Porc_Ded_Nom 1 4969.9 5029.9
## - Punt_Acierta 1 4970.6 5030.6
## - ConyLabora 1 4972.3 5032.3
## - Porc_Cap_Pago 1 4973.4 5033.4
## - NivelEduc 1 4973.7 5033.7
## - Porc_Gast_Fina 1 4975.1 5035.1
## - CateAfil 1 4975.5 5035.5
## - sexo2 1 4976.2 5036.2
## - Edad 1 4976.2 5036.2
## - TipoContrato 1 4976.3 5036.3
## - Tot_Cuot_Pag 1 4979.6 5039.6
## - Porc_Proy_Liquid_Resp_Credito 1 5019.6 5079.6
##
## Step: AIC=5025.69
## Default ~ Edad + Ingreso_Basico + Aux_Transporte + Otros_Ingresos +
## Base_70_Ingr_consti_Sal + Ded_Nomina + Tot_Gast_Fami + Cuot_Men_Tdc_Cred_Rot +
## Cuot_Mens_Otras_Oblig + Sal_Disp + Sal_Obli_Ded_Nomina +
## Sal_Créd_Viv_Veh + Sal_TDC_Cred_Rot + Punt_Acierta + Tot_Cuot_Pag +
## Porc_Proy_Liquid_Resp_Credito + Porc_Ded_Nom + Porc_Cap_Pago +
## Porc_Gast_Fina + `End_Max_ Vec_Salario` + Mont_Credito +
## CateAfil + sexo2 + ConyLabora + TipoVivienda + NivelEduc +
## TipoContrato + CumpleTiempoVincul + MoraObliVig
##
## Df Deviance AIC
## - TipoVivienda 1 4966.4 5024.4
## - Sal_Obli_Ded_Nomina 1 4966.5 5024.5
## - Aux_Transporte 1 4966.6 5024.6
## - Base_70_Ingr_consti_Sal 1 4966.7 5024.7
## - Cuot_Mens_Otras_Oblig 1 4966.7 5024.7
## - Tot_Gast_Fami 1 4966.7 5024.7
## - Otros_Ingresos 1 4966.7 5024.7
## - Sal_Disp 1 4966.7 5024.7
## - Ded_Nomina 1 4966.7 5024.7
## - Ingreso_Basico 1 4966.7 5024.7
## - Cuot_Men_Tdc_Cred_Rot 1 4966.9 5024.9
## - CumpleTiempoVincul 1 4966.9 5024.9
## - Sal_Créd_Viv_Veh 1 4967.0 5025.0
## - Sal_TDC_Cred_Rot 1 4967.5 5025.5
## - `End_Max_ Vec_Salario` 1 4967.6 5025.6
## <none> 4965.7 5025.7
## - MoraObliVig 1 4968.7 5026.7
## - Porc_Ded_Nom 1 4970.0 5028.0
## - Mont_Credito 1 4971.3 5029.3
## - Punt_Acierta 1 4971.4 5029.4
## - ConyLabora 1 4972.4 5030.4
## - Porc_Cap_Pago 1 4973.5 5031.5
## - NivelEduc 1 4974.1 5032.1
## - Porc_Gast_Fina 1 4975.1 5033.1
## - sexo2 1 4976.6 5034.6
## - CateAfil 1 4976.6 5034.6
## - TipoContrato 1 4977.0 5035.0
## - Edad 1 4977.1 5035.1
## - Tot_Cuot_Pag 1 4983.0 5041.0
## - Porc_Proy_Liquid_Resp_Credito 1 5019.7 5077.7
##
## Step: AIC=5024.39
## Default ~ Edad + Ingreso_Basico + Aux_Transporte + Otros_Ingresos +
## Base_70_Ingr_consti_Sal + Ded_Nomina + Tot_Gast_Fami + Cuot_Men_Tdc_Cred_Rot +
## Cuot_Mens_Otras_Oblig + Sal_Disp + Sal_Obli_Ded_Nomina +
## Sal_Créd_Viv_Veh + Sal_TDC_Cred_Rot + Punt_Acierta + Tot_Cuot_Pag +
## Porc_Proy_Liquid_Resp_Credito + Porc_Ded_Nom + Porc_Cap_Pago +
## Porc_Gast_Fina + `End_Max_ Vec_Salario` + Mont_Credito +
## CateAfil + sexo2 + ConyLabora + NivelEduc + TipoContrato +
## CumpleTiempoVincul + MoraObliVig
##
## Df Deviance AIC
## - Sal_Obli_Ded_Nomina 1 4967.2 5023.2
## - Aux_Transporte 1 4967.3 5023.3
## - Base_70_Ingr_consti_Sal 1 4967.4 5023.4
## - Cuot_Mens_Otras_Oblig 1 4967.4 5023.4
## - Tot_Gast_Fami 1 4967.4 5023.4
## - Otros_Ingresos 1 4967.4 5023.4
## - Sal_Disp 1 4967.4 5023.4
## - Ded_Nomina 1 4967.4 5023.4
## - Ingreso_Basico 1 4967.4 5023.4
## - Cuot_Men_Tdc_Cred_Rot 1 4967.6 5023.6
## - CumpleTiempoVincul 1 4967.6 5023.6
## - Sal_Créd_Viv_Veh 1 4967.9 5023.9
## - Sal_TDC_Cred_Rot 1 4968.2 5024.2
## - `End_Max_ Vec_Salario` 1 4968.3 5024.3
## <none> 4966.4 5024.4
## - MoraObliVig 1 4969.5 5025.5
## - Porc_Ded_Nom 1 4970.5 5026.5
## - Mont_Credito 1 4971.9 5027.9
## - Punt_Acierta 1 4972.1 5028.1
## - ConyLabora 1 4973.0 5029.0
## - Porc_Cap_Pago 1 4974.1 5030.1
## - NivelEduc 1 4974.8 5030.8
## - Porc_Gast_Fina 1 4975.7 5031.7
## - sexo2 1 4977.4 5033.4
## - CateAfil 1 4977.4 5033.4
## - TipoContrato 1 4977.8 5033.8
## - Edad 1 4978.9 5034.9
## - Tot_Cuot_Pag 1 4983.5 5039.5
## - Porc_Proy_Liquid_Resp_Credito 1 5021.0 5077.0
##
## Step: AIC=5023.18
## Default ~ Edad + Ingreso_Basico + Aux_Transporte + Otros_Ingresos +
## Base_70_Ingr_consti_Sal + Ded_Nomina + Tot_Gast_Fami + Cuot_Men_Tdc_Cred_Rot +
## Cuot_Mens_Otras_Oblig + Sal_Disp + Sal_Créd_Viv_Veh + Sal_TDC_Cred_Rot +
## Punt_Acierta + Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito +
## Porc_Ded_Nom + Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + CateAfil + sexo2 + ConyLabora + NivelEduc +
## TipoContrato + CumpleTiempoVincul + MoraObliVig
##
## Df Deviance AIC
## - Aux_Transporte 1 4968.1 5022.1
## - Base_70_Ingr_consti_Sal 1 4968.2 5022.2
## - Cuot_Mens_Otras_Oblig 1 4968.2 5022.2
## - Otros_Ingresos 1 4968.2 5022.2
## - Sal_Disp 1 4968.2 5022.2
## - Tot_Gast_Fami 1 4968.2 5022.2
## - Ded_Nomina 1 4968.2 5022.2
## - Ingreso_Basico 1 4968.2 5022.2
## - CumpleTiempoVincul 1 4968.3 5022.3
## - Cuot_Men_Tdc_Cred_Rot 1 4968.3 5022.3
## - Sal_Créd_Viv_Veh 1 4968.4 5022.4
## - Sal_TDC_Cred_Rot 1 4968.7 5022.7
## <none> 4967.2 5023.2
## - MoraObliVig 1 4970.3 5024.3
## - `End_Max_ Vec_Salario` 1 4970.7 5024.7
## - Porc_Ded_Nom 1 4971.8 5025.8
## - Punt_Acierta 1 4972.9 5026.9
## - ConyLabora 1 4973.9 5027.9
## - Mont_Credito 1 4974.3 5028.3
## - Porc_Cap_Pago 1 4974.8 5028.8
## - NivelEduc 1 4975.8 5029.8
## - Porc_Gast_Fina 1 4976.6 5030.6
## - sexo2 1 4978.2 5032.2
## - CateAfil 1 4978.4 5032.4
## - TipoContrato 1 4978.5 5032.5
## - Edad 1 4980.1 5034.1
## - Tot_Cuot_Pag 1 4984.8 5038.8
## - Porc_Proy_Liquid_Resp_Credito 1 5021.0 5075.0
##
## Step: AIC=5022.13
## Default ~ Edad + Ingreso_Basico + Otros_Ingresos + Base_70_Ingr_consti_Sal +
## Ded_Nomina + Tot_Gast_Fami + Cuot_Men_Tdc_Cred_Rot + Cuot_Mens_Otras_Oblig +
## Sal_Disp + Sal_Créd_Viv_Veh + Sal_TDC_Cred_Rot + Punt_Acierta +
## Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito + Porc_Ded_Nom +
## Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + CateAfil + sexo2 + ConyLabora + NivelEduc +
## TipoContrato + CumpleTiempoVincul + MoraObliVig
##
## Df Deviance AIC
## - CumpleTiempoVincul 1 4969.3 5021.3
## - Sal_Créd_Viv_Veh 1 4969.4 5021.4
## - Sal_TDC_Cred_Rot 1 4969.6 5021.6
## - Base_70_Ingr_consti_Sal 1 4970.0 5022.0
## <none> 4968.1 5022.1
## - Cuot_Mens_Otras_Oblig 1 4970.5 5022.5
## - Cuot_Men_Tdc_Cred_Rot 1 4970.6 5022.6
## - Tot_Gast_Fami 1 4970.9 5022.9
## - Sal_Disp 1 4971.1 5023.1
## - Otros_Ingresos 1 4971.1 5023.1
## - MoraObliVig 1 4971.3 5023.3
## - `End_Max_ Vec_Salario` 1 4971.6 5023.6
## - Ded_Nomina 1 4971.8 5023.8
## - Ingreso_Basico 1 4972.1 5024.1
## - Porc_Ded_Nom 1 4972.7 5024.7
## - Punt_Acierta 1 4973.7 5025.7
## - ConyLabora 1 4974.9 5026.9
## - Mont_Credito 1 4975.3 5027.3
## - Porc_Cap_Pago 1 4975.8 5027.8
## - NivelEduc 1 4976.8 5028.8
## - Porc_Gast_Fina 1 4977.6 5029.6
## - sexo2 1 4979.0 5031.0
## - CateAfil 1 4979.3 5031.3
## - TipoContrato 1 4979.5 5031.5
## - Edad 1 4981.1 5033.1
## - Tot_Cuot_Pag 1 4985.8 5037.8
## - Porc_Proy_Liquid_Resp_Credito 1 5022.0 5074.0
##
## Step: AIC=5021.28
## Default ~ Edad + Ingreso_Basico + Otros_Ingresos + Base_70_Ingr_consti_Sal +
## Ded_Nomina + Tot_Gast_Fami + Cuot_Men_Tdc_Cred_Rot + Cuot_Mens_Otras_Oblig +
## Sal_Disp + Sal_Créd_Viv_Veh + Sal_TDC_Cred_Rot + Punt_Acierta +
## Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito + Porc_Ded_Nom +
## Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + CateAfil + sexo2 + ConyLabora + NivelEduc +
## TipoContrato + MoraObliVig
##
## Df Deviance AIC
## - Sal_Créd_Viv_Veh 1 4970.5 5020.5
## - Sal_TDC_Cred_Rot 1 4970.8 5020.8
## - Base_70_Ingr_consti_Sal 1 4971.1 5021.1
## <none> 4969.3 5021.3
## - Cuot_Mens_Otras_Oblig 1 4971.6 5021.6
## - Cuot_Men_Tdc_Cred_Rot 1 4971.7 5021.7
## - Tot_Gast_Fami 1 4972.0 5022.0
## - Sal_Disp 1 4972.2 5022.2
## - Otros_Ingresos 1 4972.2 5022.2
## - MoraObliVig 1 4972.5 5022.5
## - `End_Max_ Vec_Salario` 1 4972.6 5022.6
## - Ded_Nomina 1 4972.9 5022.9
## - Ingreso_Basico 1 4973.2 5023.2
## - Porc_Ded_Nom 1 4974.1 5024.1
## - Punt_Acierta 1 4974.8 5024.8
## - ConyLabora 1 4976.2 5026.2
## - Mont_Credito 1 4976.9 5026.9
## - Porc_Cap_Pago 1 4977.1 5027.1
## - NivelEduc 1 4977.6 5027.6
## - Porc_Gast_Fina 1 4978.8 5028.8
## - TipoContrato 1 4979.5 5029.5
## - sexo2 1 4980.1 5030.1
## - CateAfil 1 4980.7 5030.7
## - Edad 1 4983.1 5033.1
## - Tot_Cuot_Pag 1 4987.5 5037.5
## - Porc_Proy_Liquid_Resp_Credito 1 5022.9 5072.9
##
## Step: AIC=5020.53
## Default ~ Edad + Ingreso_Basico + Otros_Ingresos + Base_70_Ingr_consti_Sal +
## Ded_Nomina + Tot_Gast_Fami + Cuot_Men_Tdc_Cred_Rot + Cuot_Mens_Otras_Oblig +
## Sal_Disp + Sal_TDC_Cred_Rot + Punt_Acierta + Tot_Cuot_Pag +
## Porc_Proy_Liquid_Resp_Credito + Porc_Ded_Nom + Porc_Cap_Pago +
## Porc_Gast_Fina + `End_Max_ Vec_Salario` + Mont_Credito +
## CateAfil + sexo2 + ConyLabora + NivelEduc + TipoContrato +
## MoraObliVig
##
## Df Deviance AIC
## - Sal_TDC_Cred_Rot 1 4972.0 5020.0
## - Base_70_Ingr_consti_Sal 1 4972.3 5020.3
## <none> 4970.5 5020.5
## - Cuot_Mens_Otras_Oblig 1 4972.9 5020.9
## - Cuot_Men_Tdc_Cred_Rot 1 4973.0 5021.0
## - Tot_Gast_Fami 1 4973.1 5021.1
## - `End_Max_ Vec_Salario` 1 4973.1 5021.1
## - Sal_Disp 1 4973.3 5021.3
## - Otros_Ingresos 1 4973.4 5021.4
## - MoraObliVig 1 4973.8 5021.8
## - Ded_Nomina 1 4974.4 5022.4
## - Ingreso_Basico 1 4974.4 5022.4
## - Porc_Ded_Nom 1 4975.7 5023.7
## - Punt_Acierta 1 4976.3 5024.3
## - Mont_Credito 1 4977.4 5025.4
## - ConyLabora 1 4977.5 5025.5
## - Porc_Cap_Pago 1 4978.2 5026.2
## - NivelEduc 1 4978.9 5026.9
## - Porc_Gast_Fina 1 4979.8 5027.8
## - TipoContrato 1 4980.8 5028.8
## - sexo2 1 4981.2 5029.2
## - CateAfil 1 4981.9 5029.9
## - Edad 1 4984.3 5032.3
## - Tot_Cuot_Pag 1 4989.4 5037.4
## - Porc_Proy_Liquid_Resp_Credito 1 5025.6 5073.6
##
## Step: AIC=5019.99
## Default ~ Edad + Ingreso_Basico + Otros_Ingresos + Base_70_Ingr_consti_Sal +
## Ded_Nomina + Tot_Gast_Fami + Cuot_Men_Tdc_Cred_Rot + Cuot_Mens_Otras_Oblig +
## Sal_Disp + Punt_Acierta + Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito +
## Porc_Ded_Nom + Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + CateAfil + sexo2 + ConyLabora + NivelEduc +
## TipoContrato + MoraObliVig
##
## Df Deviance AIC
## - Base_70_Ingr_consti_Sal 1 4973.7 5019.7
## <none> 4972.0 5020.0
## - Cuot_Men_Tdc_Cred_Rot 1 4974.2 5020.2
## - Cuot_Mens_Otras_Oblig 1 4974.3 5020.3
## - Tot_Gast_Fami 1 4974.4 5020.4
## - `End_Max_ Vec_Salario` 1 4974.6 5020.6
## - Otros_Ingresos 1 4974.7 5020.7
## - Sal_Disp 1 4974.7 5020.7
## - MoraObliVig 1 4975.3 5021.3
## - Ingreso_Basico 1 4975.8 5021.8
## - Ded_Nomina 1 4975.8 5021.8
## - Porc_Ded_Nom 1 4977.2 5023.2
## - Punt_Acierta 1 4977.7 5023.7
## - ConyLabora 1 4978.6 5024.6
## - Mont_Credito 1 4978.7 5024.7
## - Porc_Cap_Pago 1 4979.5 5025.5
## - NivelEduc 1 4980.6 5026.6
## - Porc_Gast_Fina 1 4981.0 5027.0
## - TipoContrato 1 4982.3 5028.3
## - sexo2 1 4982.6 5028.6
## - CateAfil 1 4983.7 5029.7
## - Edad 1 4985.8 5031.8
## - Tot_Cuot_Pag 1 4990.7 5036.7
## - Porc_Proy_Liquid_Resp_Credito 1 5026.7 5072.7
##
## Step: AIC=5019.71
## Default ~ Edad + Ingreso_Basico + Otros_Ingresos + Ded_Nomina +
## Tot_Gast_Fami + Cuot_Men_Tdc_Cred_Rot + Cuot_Mens_Otras_Oblig +
## Sal_Disp + Punt_Acierta + Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito +
## Porc_Ded_Nom + Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + CateAfil + sexo2 + ConyLabora + NivelEduc +
## TipoContrato + MoraObliVig
##
## Df Deviance AIC
## - Cuot_Men_Tdc_Cred_Rot 1 4974.2 5018.2
## - Tot_Gast_Fami 1 4974.4 5018.4
## - Cuot_Mens_Otras_Oblig 1 4974.5 5018.5
## <none> 4973.7 5019.7
## - Sal_Disp 1 4975.9 5019.9
## - Otros_Ingresos 1 4976.0 5020.0
## - `End_Max_ Vec_Salario` 1 4976.4 5020.4
## - MoraObliVig 1 4977.0 5021.0
## - Porc_Ded_Nom 1 4978.2 5022.2
## - ConyLabora 1 4979.4 5023.4
## - Punt_Acierta 1 4979.7 5023.7
## - Porc_Cap_Pago 1 4980.0 5024.0
## - Ded_Nomina 1 4980.1 5024.1
## - Mont_Credito 1 4980.5 5024.5
## - Porc_Gast_Fina 1 4981.5 5025.5
## - NivelEduc 1 4982.8 5026.8
## - TipoContrato 1 4983.5 5027.5
## - CateAfil 1 4983.8 5027.8
## - sexo2 1 4984.7 5028.7
## - Ingreso_Basico 1 4985.5 5029.5
## - Edad 1 4987.0 5031.0
## - Tot_Cuot_Pag 1 4991.5 5035.5
## - Porc_Proy_Liquid_Resp_Credito 1 5028.3 5072.3
##
## Step: AIC=5018.2
## Default ~ Edad + Ingreso_Basico + Otros_Ingresos + Ded_Nomina +
## Tot_Gast_Fami + Cuot_Mens_Otras_Oblig + Sal_Disp + Punt_Acierta +
## Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito + Porc_Ded_Nom +
## Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + CateAfil + sexo2 + ConyLabora + NivelEduc +
## TipoContrato + MoraObliVig
##
## Df Deviance AIC
## - Cuot_Mens_Otras_Oblig 1 4974.6 5016.6
## - Tot_Gast_Fami 1 4974.9 5016.9
## - Sal_Disp 1 4975.9 5017.9
## - Otros_Ingresos 1 4976.0 5018.0
## <none> 4974.2 5018.2
## - `End_Max_ Vec_Salario` 1 4976.9 5018.9
## - MoraObliVig 1 4977.6 5019.6
## - Porc_Ded_Nom 1 4979.1 5021.1
## - Ded_Nomina 1 4980.1 5022.1
## - Punt_Acierta 1 4980.2 5022.2
## - ConyLabora 1 4980.6 5022.6
## - Porc_Cap_Pago 1 4980.6 5022.6
## - Mont_Credito 1 4980.9 5022.9
## - Porc_Gast_Fina 1 4982.0 5024.0
## - NivelEduc 1 4983.2 5025.2
## - TipoContrato 1 4984.0 5026.0
## - CateAfil 1 4984.5 5026.5
## - sexo2 1 4985.1 5027.1
## - Ingreso_Basico 1 4985.7 5027.7
## - Edad 1 4987.5 5029.5
## - Tot_Cuot_Pag 1 4992.3 5034.3
## - Porc_Proy_Liquid_Resp_Credito 1 5028.8 5070.8
##
## Step: AIC=5016.59
## Default ~ Edad + Ingreso_Basico + Otros_Ingresos + Ded_Nomina +
## Tot_Gast_Fami + Sal_Disp + Punt_Acierta + Tot_Cuot_Pag +
## Porc_Proy_Liquid_Resp_Credito + Porc_Ded_Nom + Porc_Cap_Pago +
## Porc_Gast_Fina + `End_Max_ Vec_Salario` + Mont_Credito +
## CateAfil + sexo2 + ConyLabora + NivelEduc + TipoContrato +
## MoraObliVig
##
## Df Deviance AIC
## - Tot_Gast_Fami 1 4975.1 5015.1
## - Otros_Ingresos 1 4976.4 5016.4
## <none> 4974.6 5016.6
## - Sal_Disp 1 4976.6 5016.6
## - `End_Max_ Vec_Salario` 1 4977.3 5017.3
## - MoraObliVig 1 4977.9 5017.9
## - Ded_Nomina 1 4980.2 5020.2
## - Punt_Acierta 1 4980.5 5020.5
## - Porc_Ded_Nom 1 4980.6 5020.6
## - Mont_Credito 1 4981.2 5021.2
## - Porc_Cap_Pago 1 4981.5 5021.5
## - ConyLabora 1 4982.1 5022.1
## - Porc_Gast_Fina 1 4982.7 5022.7
## - NivelEduc 1 4983.4 5023.4
## - TipoContrato 1 4984.5 5024.5
## - CateAfil 1 4985.3 5025.3
## - sexo2 1 4985.3 5025.3
## - Edad 1 4988.0 5028.0
## - Ingreso_Basico 1 4991.7 5031.7
## - Tot_Cuot_Pag 1 4993.9 5033.9
## - Porc_Proy_Liquid_Resp_Credito 1 5029.0 5069.0
##
## Step: AIC=5015.14
## Default ~ Edad + Ingreso_Basico + Otros_Ingresos + Ded_Nomina +
## Sal_Disp + Punt_Acierta + Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito +
## Porc_Ded_Nom + Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + CateAfil + sexo2 + ConyLabora + NivelEduc +
## TipoContrato + MoraObliVig
##
## Df Deviance AIC
## - Otros_Ingresos 1 4976.6 5014.6
## - Sal_Disp 1 4976.9 5014.9
## <none> 4975.1 5015.1
## - `End_Max_ Vec_Salario` 1 4977.9 5015.9
## - MoraObliVig 1 4978.5 5016.5
## - Ded_Nomina 1 4981.1 5019.1
## - Punt_Acierta 1 4981.2 5019.2
## - Mont_Credito 1 4981.9 5019.9
## - ConyLabora 1 4982.2 5020.2
## - Porc_Ded_Nom 1 4982.6 5020.6
## - Porc_Cap_Pago 1 4983.6 5021.6
## - NivelEduc 1 4983.9 5021.9
## - TipoContrato 1 4985.0 5023.0
## - Porc_Gast_Fina 1 4985.3 5023.3
## - sexo2 1 4985.8 5023.8
## - CateAfil 1 4987.0 5025.0
## - Edad 1 4988.7 5026.7
## - Ingreso_Basico 1 4994.0 5032.0
## - Tot_Cuot_Pag 1 4997.2 5035.2
## - Porc_Proy_Liquid_Resp_Credito 1 5030.2 5068.2
##
## Step: AIC=5014.58
## Default ~ Edad + Ingreso_Basico + Ded_Nomina + Sal_Disp + Punt_Acierta +
## Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito + Porc_Ded_Nom +
## Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + CateAfil + sexo2 + ConyLabora + NivelEduc +
## TipoContrato + MoraObliVig
##
## Df Deviance AIC
## - Sal_Disp 1 4977.1 5013.1
## <none> 4976.6 5014.6
## - `End_Max_ Vec_Salario` 1 4979.4 5015.4
## - MoraObliVig 1 4979.9 5015.9
## - Ded_Nomina 1 4981.9 5017.9
## - Punt_Acierta 1 4982.9 5018.9
## - Mont_Credito 1 4983.8 5019.8
## - NivelEduc 1 4985.2 5021.2
## - ConyLabora 1 4986.1 5022.1
## - Porc_Ded_Nom 1 4986.1 5022.1
## - TipoContrato 1 4986.4 5022.4
## - Porc_Cap_Pago 1 4987.1 5023.1
## - sexo2 1 4987.5 5023.5
## - CateAfil 1 4988.9 5024.9
## - Edad 1 4990.1 5026.1
## - Porc_Gast_Fina 1 4990.1 5026.1
## - Ingreso_Basico 1 4994.6 5030.6
## - Tot_Cuot_Pag 1 5002.1 5038.1
## - Porc_Proy_Liquid_Resp_Credito 1 5030.9 5066.9
##
## Step: AIC=5013.13
## Default ~ Edad + Ingreso_Basico + Ded_Nomina + Punt_Acierta +
## Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito + Porc_Ded_Nom +
## Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + CateAfil + sexo2 + ConyLabora + NivelEduc +
## TipoContrato + MoraObliVig
##
## Df Deviance AIC
## <none> 4977.1 5013.1
## - `End_Max_ Vec_Salario` 1 4979.7 5013.7
## - MoraObliVig 1 4980.5 5014.5
## - Ded_Nomina 1 4982.0 5016.0
## - Punt_Acierta 1 4983.7 5017.7
## - Mont_Credito 1 4984.1 5018.1
## - NivelEduc 1 4986.1 5020.1
## - Porc_Ded_Nom 1 4986.6 5020.6
## - ConyLabora 1 4987.0 5021.0
## - TipoContrato 1 4987.1 5021.1
## - Porc_Cap_Pago 1 4987.4 5021.4
## - sexo2 1 4987.9 5021.9
## - CateAfil 1 4990.8 5024.8
## - Edad 1 4990.9 5024.9
## - Porc_Gast_Fina 1 4991.4 5025.4
## - Ingreso_Basico 1 4994.6 5028.6
## - Tot_Cuot_Pag 1 5005.0 5039.0
## - Porc_Proy_Liquid_Resp_Credito 1 5031.3 5065.3mod_2##
## Call: glm(formula = Default ~ Edad + Ingreso_Basico + Ded_Nomina +
## Punt_Acierta + Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito +
## Porc_Ded_Nom + Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + CateAfil + sexo2 + ConyLabora + NivelEduc +
## TipoContrato + MoraObliVig, family = "binomial", data = datos)
##
## Coefficients:
## (Intercept) Edad
## 1.921e+00 -1.696e-02
## Ingreso_Basico Ded_Nomina
## 6.821e-07 -8.242e-07
## Punt_Acierta Tot_Cuot_Pag
## -5.218e-04 -7.462e-06
## Porc_Proy_Liquid_Resp_Credito Porc_Ded_Nom
## -4.043e-01 4.033e+00
## Porc_Cap_Pago Porc_Gast_Fina
## -5.558e+00 6.378e+00
## `End_Max_ Vec_Salario` Mont_Credito
## -4.140e-02 7.578e-08
## CateAfil sexo2
## -3.502e-01 2.972e-01
## ConyLabora NivelEduc
## -8.900e-01 -1.340e-01
## TipoContrato MoraObliVig
## -5.176e-02 -3.197e-02
##
## Degrees of Freedom: 15812 Total (i.e. Null); 15795 Residual
## Null Deviance: 5251
## Residual Deviance: 4977 AIC: 5013summary(mod_2)##
## Call:
## glm(formula = Default ~ Edad + Ingreso_Basico + Ded_Nomina +
## Punt_Acierta + Tot_Cuot_Pag + Porc_Proy_Liquid_Resp_Credito +
## Porc_Ded_Nom + Porc_Cap_Pago + Porc_Gast_Fina + `End_Max_ Vec_Salario` +
## Mont_Credito + CateAfil + sexo2 + ConyLabora + NivelEduc +
## TipoContrato + MoraObliVig, family = "binomial", data = datos)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.4089 -0.3227 -0.2608 -0.1944 3.5998
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.921e+00 8.131e-01 2.363 0.018130 *
## Edad -1.696e-02 4.660e-03 -3.640 0.000273 ***
## Ingreso_Basico 6.821e-07 1.546e-07 4.411 1.03e-05 ***
## Ded_Nomina -8.242e-07 4.001e-07 -2.060 0.039379 *
## Punt_Acierta -5.218e-04 1.979e-04 -2.637 0.008367 **
## Tot_Cuot_Pag -7.462e-06 1.507e-06 -4.953 7.31e-07 ***
## Porc_Proy_Liquid_Resp_Credito -4.043e-01 6.601e-02 -6.125 9.06e-10 ***
## Porc_Ded_Nom 4.033e+00 1.293e+00 3.118 0.001818 **
## Porc_Cap_Pago -5.558e+00 1.715e+00 -3.241 0.001193 **
## Porc_Gast_Fina 6.378e+00 1.664e+00 3.832 0.000127 ***
## `End_Max_ Vec_Salario` -4.140e-02 2.628e-02 -1.576 0.115140
## Mont_Credito 7.578e-08 2.886e-08 2.626 0.008652 **
## CateAfil -3.502e-01 9.438e-02 -3.711 0.000207 ***
## sexo2 2.972e-01 9.133e-02 3.254 0.001136 **
## ConyLabora -8.900e-01 2.810e-01 -3.167 0.001542 **
## NivelEduc -1.340e-01 4.513e-02 -2.968 0.002994 **
## TipoContrato -5.176e-02 1.632e-02 -3.171 0.001518 **
## MoraObliVig -3.197e-02 1.691e-02 -1.890 0.058749 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5250.7 on 15812 degrees of freedom
## Residual deviance: 4977.1 on 15795 degrees of freedom
## AIC: 5013.1
##
## Number of Fisher Scoring iterations: 72.3. Usando Caret y Cross Validation para las variables
require(caret)## Loading required package: caret## Loading required package: ggplot2## Loading required package: lattice2.3.1. Modelo de Regresión logística
control=trainControl(method = "cv",number = 10)
mod_caret=train(Default~.,data=datos,method="glm",family="binomial",trControl=control,metric="Accuracy")varImp(mod_caret)## glm variable importance
##
## only 20 most important variables shown (out of 35)
##
## Overall
## Porc_Proy_Liquid_Resp_Credito 100.00
## Tot_Cuot_Pag 84.31
## TipoContrato 75.30
## sexo2 72.79
## Porc_Gast_Fina 70.67
## CateAfil 70.03
## Edad 67.45
## Porc_Cap_Pago 62.75
## NivelEduc 59.09
## ConyLabora 57.05
## Punt_Acierta 52.16
## Mont_Credito 48.55
## Porc_Ded_Nom 47.89
## MoraObliVig 41.17
## Sal_TDC_Cred_Rot 32.19
## `\\`End_Max_ Vec_Salario\\`` 28.61
## CumpleTiempoVincul 26.41
## TipoVivienda 19.92
## Sal_Créd_Viv_Veh 17.88
## Sal_Obli_Ded_Nomina 17.53plot(varImp(mod_caret))
most_impo=varImp(mod_caret)
most_impo## glm variable importance
##
## only 20 most important variables shown (out of 35)
##
## Overall
## Porc_Proy_Liquid_Resp_Credito 100.00
## Tot_Cuot_Pag 84.31
## TipoContrato 75.30
## sexo2 72.79
## Porc_Gast_Fina 70.67
## CateAfil 70.03
## Edad 67.45
## Porc_Cap_Pago 62.75
## NivelEduc 59.09
## ConyLabora 57.05
## Punt_Acierta 52.16
## Mont_Credito 48.55
## Porc_Ded_Nom 47.89
## MoraObliVig 41.17
## Sal_TDC_Cred_Rot 32.19
## `\\`End_Max_ Vec_Salario\\`` 28.61
## CumpleTiempoVincul 26.41
## TipoVivienda 19.92
## Sal_Créd_Viv_Veh 17.88
## Sal_Obli_Ded_Nomina 17.53Variables Top
datos=data.frame(datos)
mod_top=glm(Default~Porc_Proy_Liquid_Resp_Credito+Tot_Cuot_Pag+TipoContrato+sexo2+Porc_Gast_Fina+CateAfil+Edad+Porc_Cap_Pago+NivelEduc+ConyLabora+Punt_Acierta+Mont_Credito+Porc_Ded_Nom+MoraObliVig,data=datos,family = "binomial")
res_coef=matrix(NA,nrow = 1000,ncol = 14)
for(i in 1:1000){
unos=which(datos$Default==1)
ceros=sample(which(datos$Default==0),623)
datos_balanceo=datos[c(unos,ceros),]
mod_top_balanceo=glm(Default~Porc_Proy_Liquid_Resp_Credito+Tot_Cuot_Pag+TipoContrato+sexo2+Porc_Gast_Fina+CateAfil+Edad+Porc_Cap_Pago+NivelEduc+ConyLabora+Punt_Acierta+Mont_Credito+Porc_Ded_Nom,data=datos_balanceo,family = "binomial")
res_coef[i,]=round(mod_top_balanceo$coefficients,8)
}res_coef=data.frame(res_coef)
names(res_coef)=names(mod_top_balanceo$coefficients)summary(mod_top_balanceo)##
## Call:
## glm(formula = Default ~ Porc_Proy_Liquid_Resp_Credito + Tot_Cuot_Pag +
## TipoContrato + sexo2 + Porc_Gast_Fina + CateAfil + Edad +
## Porc_Cap_Pago + NivelEduc + ConyLabora + Punt_Acierta + Mont_Credito +
## Porc_Ded_Nom, family = "binomial", data = datos_balanceo)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.8329 -1.1376 0.2915 1.0867 2.3904
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 3.668e+00 9.934e-01 3.692 0.000223 ***
## Porc_Proy_Liquid_Resp_Credito -3.377e-01 7.953e-02 -4.246 2.18e-05 ***
## Tot_Cuot_Pag -2.899e-06 1.529e-06 -1.896 0.058004 .
## TipoContrato -2.312e-02 2.338e-02 -0.989 0.322714
## sexo2 2.493e-01 1.293e-01 1.928 0.053884 .
## Porc_Gast_Fina 2.827e+00 2.119e+00 1.334 0.182211
## CateAfil -1.788e-01 1.285e-01 -1.391 0.164286
## Edad -1.374e-02 6.315e-03 -2.175 0.029632 *
## Porc_Cap_Pago -2.745e+00 2.214e+00 -1.240 0.215076
## NivelEduc -4.967e-02 6.248e-02 -0.795 0.426605
## ConyLabora -5.039e-01 3.673e-01 -1.372 0.170145
## Punt_Acierta -9.139e-04 3.258e-04 -2.805 0.005025 **
## Mont_Credito 3.361e-08 2.882e-08 1.166 0.243513
## Porc_Ded_Nom 3.045e-01 1.554e+00 0.196 0.844599
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 1727.3 on 1245 degrees of freedom
## Residual deviance: 1618.5 on 1232 degrees of freedom
## AIC: 1646.5
##
## Number of Fisher Scoring iterations: 43. Etapa 3: Balanceo de datos Oversampling
require(reticulate)# Importamos la librerías
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
#%matplotlib inline
import seaborn as sns
import warnings
warnings.filterwarnings('ignore')
from imblearn.over_sampling import SMOTE
from imblearn.pipeline import make_pipeline
from pylab import rcParams
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import accuracy_score
from sklearn.metrics import precision_score, recall_score
from sklearn.metrics import f1_score, roc_auc_score, roc_curve
from sklearn.model_selection import train_test_split
from sklearn.model_selection import GridSearchCVCon el método de reducción de dimensionalidad Caret y Cross Validation para las variables, seleccionamos las variables más importantes:
Default = Porc_Proy_Liquid_Resp_Credito + Tot_Cuot_Pag + TipoContrato + sexo2 + Porc_Gast_Fina + CateAfil + Edad + Porc_Cap_Pago + NivelEduc + ConyLabora + Punt_Acierta + Mont_Credito + Porc_Ded_Nom
df = pd.read_excel("D:/01_PUJ/00_Trabajo_de_grado/Data/DatosBD27052023RedDim3.xlsx")df.head()## Edad Punt_Acierta Tot_Cuot_Pag ... NivelEduc TipoContrato Default
## 0 40 683 96516.136875 ... 2 7 0
## 1 57 803 72842.367453 ... 2 7 0
## 2 31 838 107736.180180 ... 3 7 0
## 3 52 665 116714.195195 ... 5 7 0
## 4 33 794 100158.255248 ... 2 7 0
##
## [5 rows x 13 columns]df.shape## (15813, 13)df.describe()## Edad Punt_Acierta ... TipoContrato Default
## count 15813.000000 15813.000000 ... 15813.000000 15813.000000
## mean 35.717764 683.636438 ... 5.323025 0.039398
## std 10.604924 186.269795 ... 2.449206 0.194546
## min 18.000000 0.000000 ... 1.000000 0.000000
## 25% 27.000000 667.000000 ... 2.000000 0.000000
## 50% 34.000000 727.000000 ... 7.000000 0.000000
## 75% 43.000000 781.000000 ... 7.000000 0.000000
## max 69.000000 938.000000 ... 7.000000 1.000000
##
## [8 rows x 13 columns]Se hace tratamiento de valores faltantes
df_missing_values = df.isnull().sum()
df_missing_values ## Edad 0
## Punt_Acierta 0
## Tot_Cuot_Pag 0
## Porc_Proy_Liquid_Resp_Credito 0
## Porc_Cap_Pago 0
## Porc_Gast_Fina 0
## Mont_Credito 0
## CateAfil 0
## sexo2 0
## ConyLabora 0
## NivelEduc 0
## TipoContrato 0
## Default 0
## dtype: int64Análisis de distribución de datos
classes = df['Default'].value_counts()
classes## 0 15190
## 1 623
## Name: Default, dtype: int64# Creamos marco de datos de créditos "En Mora"
data_EnMora = df[df['Default'] == 1]
# Creamos marco de datos de créditos "Al Dia"
data_AlDia = df[df['Default'] == 0]Dividimos en Entrenamiento y Test
# Importamos librería
from sklearn.model_selection import train_test_split# Introducir variables de características en X
X = df.drop(['Default'], axis=1)# Poniendo la variable objetivo en y
y = df['Default']# División de datos en Entrenamiento y Test de prueba 80:20
X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=0.8, test_size=0.2, random_state=100)Balanceo de datos son el Método SMOTE OVERSAMPLING
unique, count = np.unique(y_train, return_counts=True)
Y_train_dict_value_count = { k:v for (k,v) in zip(unique, count)}
Y_train_dict_value_count## {0: 12148, 1: 502}sm = SMOTE(random_state=12, sampling_strategy = 1.0)
x_train_res, y_train_res = sm.fit_resample(X_train, y_train)unique, count = np.unique(y_train_res, return_counts=True)
y_train_smote_value_count = { k:v for (k,v) in zip(unique, count)}
y_train_smote_value_count## {0: 12148, 1: 12148}4. Etapa 4: Modelado
Dividir el conjunto de datos en conjuntos de entrenamiento y test.
Realizar el entrenamiento del modelo de regresión logística utilizando para ello el conjunto de entrenamiento y compararlo con los modelos XGBoost, árbol de decisión, Bosque aleatorio.
Aplicar el modelo construido a los datos del conjunto de test.
Cálculo de una métrica de desempeño para evaluar la calidad del modelo en sus prediciones sobre el conjunto de test.
# Importando métricas
from sklearn import metrics
from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay, classification_report, f1_scoreresults = pd.DataFrame(columns=['Model Name', 'Accuracy', 'F1-score', 'ROC'])# Función de curva ROC
def draw_roc( actual, probs ):
fpr, tpr, thresholds = metrics.roc_curve( actual, probs,
drop_intermediate = False )
auc_score = metrics.roc_auc_score( actual, probs )
plt.figure(figsize=(5, 5))
plt.plot( fpr, tpr, label='ROC curve (area = %0.2f)' % auc_score )
plt.plot([0, 1], [0, 1], 'k--')
plt.xlim([0.0, 1.0])
plt.ylim([0.0, 1.05])
plt.xlabel('False Positive Rate or [1 - True Negative Rate]')
plt.ylabel('True Positive Rate')
plt.title('Receiver operating characteristic example')
plt.legend(loc="lower right")
plt.show()
return None4.1. Modelo de Regresión Logística
# Importando el módulo de regresión logística de scikit
from sklearn.linear_model import LogisticRegression# Instanciar el modelo con mejor C
logistic = LogisticRegression(C=0.01)# Ajuste el modelo en el set de entrenamiento
logistic_model = logistic.fit(x_train_res, y_train_res)# Preparar función de resultados
def display_test_results(model_name, model):
# Prediction on the test set
y_test_pred = model.predict(X_test)
# Confusion matrix
print("------------------ Confusion Matrix --------------------")
c_matrix = metrics.confusion_matrix(y_test, y_test_pred)
print(c_matrix)
cm_display = ConfusionMatrixDisplay(confusion_matrix=c_matrix)
cm_display.plot(cmap=plt.cm.Blues)
plt.show()
# classification_report
print("------------------ classification_report --------------------")
print(classification_report(y_test, y_test_pred))
print("------------------ More Specific classification_report --------------------")
TP = c_matrix[1,1] # true positive
TN = c_matrix[0,0] # true negatives
FP = c_matrix[0,1] # false positives
FN = c_matrix[1,0] # false negatives
# Accuracy
print("Accuracy:-",metrics.accuracy_score(y_test, y_test_pred))
# Sensitivity
print("Sensitivity:-",TP / float(TP+FN))
# Specificity
print("Specificity:-", TN / float(TN+FP))
# F1 score
print("F1-Score:-", f1_score(y_test, y_test_pred))
# Predicted probability
y_test_pred_proba = model.predict_proba(X_test)[:,1]
# roc_auc
print("------------------ ROC --------------------")
roc_auc = metrics.roc_auc_score(y_test, y_test_pred_proba)
# Plot the ROC curve
draw_roc(y_test, y_test_pred_proba)
# add all metrics score in final result store
results.loc[len(results)] = [model_name, metrics.accuracy_score(y_test, y_test_pred), f1_score(y_test, y_test_pred), roc_auc]
return None4.1.1. Resultados de predicción Modelo de Regresión Logística
display_test_results("Logistic Regression", logistic_model)## ------------------ Confusion Matrix --------------------
## [[3042 0]
## [ 121 0]]
## ------------------ classification_report --------------------
## precision recall f1-score support
##
## 0 0.96 1.00 0.98 3042
## 1 0.00 0.00 0.00 121
##
## accuracy 0.96 3163
## macro avg 0.48 0.50 0.49 3163
## weighted avg 0.92 0.96 0.94 3163
##
## ------------------ More Specific classification_report --------------------
## Accuracy:- 0.9617451786278849
## Sensitivity:- 0.0
## Specificity:- 1.0
## F1-Score:- 0.0
## ------------------ ROC --------------------
4.2. Modelo XGBoost
# Importamos XGBoost
from xgboost import XGBClassifierparams = {'learning_rate': 0.2,
'max_depth': 2,
'n_estimators':200,
'subsample':0.9,
'objective':'binary:logistic'}# modelo de ajuste en datos de entrenamiento
xgb_model = XGBClassifier(params = params)
xgb_model.fit(x_train_res, y_train_res)## [18:50:17] WARNING: C:/Users/administrator/workspace/xgboost-win64_release_1.6.0/src/learner.cc:627:
## Parameters: { "params" } might not be used.
##
## This could be a false alarm, with some parameters getting used by language bindings but
## then being mistakenly passed down to XGBoost core, or some parameter actually being used
## but getting flagged wrongly here. Please open an issue if you find any such cases.
##
##
## XGBClassifier(base_score=0.5, booster='gbtree', callbacks=None,
## colsample_bylevel=1, colsample_bynode=1, colsample_bytree=1,
## early_stopping_rounds=None, enable_categorical=False,
## eval_metric=None, gamma=0, gpu_id=-1, grow_policy='depthwise',
## importance_type=None, interaction_constraints='',
## learning_rate=0.300000012, max_bin=256, max_cat_to_onehot=4,
## max_delta_step=0, max_depth=6, max_leaves=0, min_child_weight=1,
## missing=nan, monotone_constraints='()', n_estimators=100,
## n_jobs=0, num_parallel_tree=1,
## params={'learning_rate': 0.2, 'max_depth': 2, 'n_estimators': 200,
## 'objective': 'binary:logistic', 'subsample': 0.9},
## predictor='auto', random_state=0, reg_alpha=0, ...)4.2.1. Resultados de predicción Modelo XGBoost
display_test_results("XG Boost", xgb_model)## ------------------ Confusion Matrix --------------------
## [[2819 223]
## [ 101 20]]
## ------------------ classification_report --------------------
## precision recall f1-score support
##
## 0 0.97 0.93 0.95 3042
## 1 0.08 0.17 0.11 121
##
## accuracy 0.90 3163
## macro avg 0.52 0.55 0.53 3163
## weighted avg 0.93 0.90 0.91 3163
##
## ------------------ More Specific classification_report --------------------
## Accuracy:- 0.89756560227632
## Sensitivity:- 0.1652892561983471
## Specificity:- 0.9266929651545036
## F1-Score:- 0.10989010989010989
## ------------------ ROC --------------------
4.3. Modelo Árbol de decisión
# Importando clasificador del árbol de decisión
from sklearn.tree import DecisionTreeClassifier# Modelo con hiperparámetros óptimos
decision_tree_model = DecisionTreeClassifier(criterion = "gini",
random_state = 100,
max_depth=5,
min_samples_leaf=100,
min_samples_split=100)
decision_tree_model.fit(x_train_res, y_train_res)## DecisionTreeClassifier(max_depth=5, min_samples_leaf=100, min_samples_split=100,
## random_state=100)4.3.1. Resultados de predicción Modelo Árbol de decisión
display_test_results("Decision Tree", decision_tree_model)## ------------------ Confusion Matrix --------------------
## [[2550 492]
## [ 82 39]]
## ------------------ classification_report --------------------
## precision recall f1-score support
##
## 0 0.97 0.84 0.90 3042
## 1 0.07 0.32 0.12 121
##
## accuracy 0.82 3163
## macro avg 0.52 0.58 0.51 3163
## weighted avg 0.93 0.82 0.87 3163
##
## ------------------ More Specific classification_report --------------------
## Accuracy:- 0.8185267151438508
## Sensitivity:- 0.32231404958677684
## Specificity:- 0.8382642998027613
## F1-Score:- 0.1196319018404908
## ------------------ ROC --------------------
4.4. Modelo Bosque aleatorio
# Importando clasificador de bosque aleatorio
from sklearn.ensemble import RandomForestClassifierrandom_forest_model = RandomForestClassifier(bootstrap=True,
max_depth=5,
min_samples_leaf=50,
min_samples_split=50,
max_features=10,
n_estimators=100)# Ajustar el modelo
random_forest_model.fit(x_train_res, y_train_res)## RandomForestClassifier(max_depth=5, max_features=10, min_samples_leaf=50,
## min_samples_split=50)4.4.1. Resultados de predicción Modelo Bosque aleatorio
display_test_results("Random Forest", random_forest_model)## ------------------ Confusion Matrix --------------------
## [[2411 631]
## [ 78 43]]
## ------------------ classification_report --------------------
## precision recall f1-score support
##
## 0 0.97 0.79 0.87 3042
## 1 0.06 0.36 0.11 121
##
## accuracy 0.78 3163
## macro avg 0.52 0.57 0.49 3163
## weighted avg 0.93 0.78 0.84 3163
##
## ------------------ More Specific classification_report --------------------
## Accuracy:- 0.7758457160923175
## Sensitivity:- 0.35537190082644626
## Specificity:- 0.7925706771860618
## F1-Score:- 0.10817610062893081
## ------------------ ROC --------------------
5. Resumen
results.sort_values(by="ROC", ascending=False)## Model Name Accuracy F1-score ROC
## 1 XG Boost 0.897566 0.109890 0.649472
## 3 Random Forest 0.775846 0.108176 0.620794
## 2 Decision Tree 0.818527 0.119632 0.608542
## 0 Logistic Regression 0.961745 0.000000 0.561807