CLASE-11
Dilmer Sandoval
2024-01
Aplicaciones
# Cargar el conjunto de datos
telco <- read.csv("https://raw.githubusercontent.com/VictorGuevaraP/Estadistica-R/master/Caso_telefon%C3%ADa.csv",
encoding = "latin1", sep = ";", stringsAsFactors = T)
head(telco)# Original
hist(telco$Edad, 12)
Para sacar la raiz cuadrada, simplemente se puede utilizar la funcion sqrt
sqrt(telco$Edad)## [1] 5.196152 5.291503 5.291503 4.582576 5.385165 5.099020 6.324555 6.324555
## [9] 5.477226 5.196152 5.291503 4.898979 5.099020 5.744563 4.582576 6.324555
## [17] 5.000000 6.164414 5.916080 5.567764 5.291503 5.099020 5.099020 5.744563
## [25] 5.830952 4.690416 5.099020 6.244998 5.567764 4.690416 5.291503 5.916080
## [33] 4.690416 6.324555 5.830952 4.472136 5.830952 4.472136 5.385165 5.744563
## [41] 4.795832 5.916080 4.898979 5.656854 5.656854 5.196152 6.082763 5.656854
## [49] 5.744563 5.916080 6.244998 5.830952 6.164414 6.244998 6.082763 5.656854
## [57] 6.082763 5.000000 5.099020 6.324555 6.324555 5.477226 5.291503 4.898979
## [65] 5.744563 4.582576 5.291503 4.582576 5.385165 6.000000 5.830952 4.472136
## [73] 5.830952 4.472136 5.385165 5.567764 4.690416 6.164414 6.000000 5.744563Graficamente
hist(sqrt(telco$Edad))
Transformacion exponencial
en r para poder obtener esta transformacion, se debe utilizar la funcion exp()
exp(telco$Edad)## [1] 5.320482e+11 1.446257e+12 1.446257e+12 1.318816e+09 3.931334e+12
## [6] 1.957296e+11 2.353853e+17 2.353853e+17 1.068647e+13 5.320482e+11
## [11] 1.446257e+12 2.648912e+10 1.957296e+11 2.146436e+14 1.318816e+09
## [16] 2.353853e+17 7.200490e+10 3.185593e+16 1.586013e+15 2.904885e+13
## [21] 1.446257e+12 1.957296e+11 1.957296e+11 2.146436e+14 5.834617e+14
## [26] 3.584913e+09 1.957296e+11 8.659340e+16 2.904885e+13 3.584913e+09
## [31] 1.446257e+12 1.586013e+15 3.584913e+09 2.353853e+17 5.834617e+14
## [36] 4.851652e+08 5.834617e+14 4.851652e+08 3.931334e+12 2.146436e+14
## [41] 9.744803e+09 1.586013e+15 2.648912e+10 7.896296e+13 7.896296e+13
## [46] 5.320482e+11 1.171914e+16 7.896296e+13 2.146436e+14 1.586013e+15
## [51] 8.659340e+16 5.834617e+14 3.185593e+16 8.659340e+16 1.171914e+16
## [56] 7.896296e+13 1.171914e+16 7.200490e+10 1.957296e+11 2.353853e+17
## [61] 2.353853e+17 1.068647e+13 1.446257e+12 2.648912e+10 2.146436e+14
## [66] 1.318816e+09 1.446257e+12 1.318816e+09 3.931334e+12 4.311232e+15
## [71] 5.834617e+14 4.851652e+08 5.834617e+14 4.851652e+08 3.931334e+12
## [76] 2.904885e+13 3.584913e+09 3.185593e+16 4.311232e+15 2.146436e+14Para poder observarlo graficamente
hist(exp(telco$Edad))
Forma 2
edad_exp<- exp(telco$Edad)
hist(edad_exp)
Tranformacion logaritmica
Para la transformacion logaritmica
log(telco$Edad)## [1] 3.295837 3.332205 3.332205 3.044522 3.367296 3.258097 3.688879 3.688879
## [9] 3.401197 3.295837 3.332205 3.178054 3.258097 3.496508 3.044522 3.688879
## [17] 3.218876 3.637586 3.555348 3.433987 3.332205 3.258097 3.258097 3.496508
## [25] 3.526361 3.091042 3.258097 3.663562 3.433987 3.091042 3.332205 3.555348
## [33] 3.091042 3.688879 3.526361 2.995732 3.526361 2.995732 3.367296 3.496508
## [41] 3.135494 3.555348 3.178054 3.465736 3.465736 3.295837 3.610918 3.465736
## [49] 3.496508 3.555348 3.663562 3.526361 3.637586 3.663562 3.610918 3.465736
## [57] 3.610918 3.218876 3.258097 3.688879 3.688879 3.401197 3.332205 3.178054
## [65] 3.496508 3.044522 3.332205 3.044522 3.367296 3.583519 3.526361 2.995732
## [73] 3.526361 2.995732 3.367296 3.433987 3.091042 3.637586 3.583519 3.496508Graficamente
hist(log(telco$Edad))
Cambiar la base 2
log(telco$Edad, base=2)## [1] 4.754888 4.807355 4.807355 4.392317 4.857981 4.700440 5.321928 5.321928
## [9] 4.906891 4.754888 4.807355 4.584963 4.700440 5.044394 4.392317 5.321928
## [17] 4.643856 5.247928 5.129283 4.954196 4.807355 4.700440 4.700440 5.044394
## [25] 5.087463 4.459432 4.700440 5.285402 4.954196 4.459432 4.807355 5.129283
## [33] 4.459432 5.321928 5.087463 4.321928 5.087463 4.321928 4.857981 5.044394
## [41] 4.523562 5.129283 4.584963 5.000000 5.000000 4.754888 5.209453 5.000000
## [49] 5.044394 5.129283 5.285402 5.087463 5.247928 5.285402 5.209453 5.000000
## [57] 5.209453 4.643856 4.700440 5.321928 5.321928 4.906891 4.807355 4.584963
## [65] 5.044394 4.392317 4.807355 4.392317 4.857981 5.169925 5.087463 4.321928
## [73] 5.087463 4.321928 4.857981 4.954196 4.459432 5.247928 5.169925 5.044394Graficamente
hist(log(telco$Edad, base=2))
Comparacion de transformaciones
# Obtener solo transformaciones
edad_sqrt <- sqrt(telco$Edad)
edad_exp <- exp(telco$Edad)
edad_ln <- log(telco$Edad)
edad_log2 <- log(telco$Edad, base=2)
edad_log5 <- log(telco$Edad, base=5)Ver graficamente
par(mfrow=c(3,2))
plot(density(telco$Edad))
plot(density(edad_sqrt))
plot(density(edad_exp))
plot(density(edad_ln))
plot(density(edad_log2))
plot(density(edad_log5))
par(mfrow=c(1,1))
hist(telco$Edad)
hist(edad_sqrt)
hist(edad_exp)
hist(edad_ln)
hist(edad_log2)
hist(edad_log5)
par(mfrow=c(1,1))Este tipo de transformaciones, busca que los datos sean simetricos (distibucion forma de campana) o que se cumpla un supuesto estadistico
Lavisualizacion de la transformacion puede mejorarse con la grafica de densidad
density(telco$Edad)##
## Call:
## density.default(x = telco$Edad)
##
## Data: telco$Edad (80 obs.); Bandwidth 'bw' = 2.254
##
## x y
## Min. :13.24 Min. :0.0001285
## 1st Qu.:21.62 1st Qu.:0.0092196
## Median :30.00 Median :0.0351818
## Mean :30.00 Mean :0.0297968
## 3rd Qu.:38.38 3rd Qu.:0.0494093
## Max. :46.76 Max. :0.0531821par(mfrow=c(3,2))
plot(density(telco$Edad), main = "Distribucion de edades originales")
plot(density(edad_sqrt), main = "Disatribucion de edades transformadas - sqrt")
plot(density(edad_exp), main = "Disatribucion de edades transformadas - exp")
plot(density(edad_ln), main = "Disatribucion de edades transformadas - log")
plot(density(edad_log2), main = "Disatribucion de edades transformadas - log2")
plot(density(edad_log5), main = "Disatribucion de edades transformadas - log5")
par(mfrow=c(1,1))library(PerformanceAnalytics)## Loading required package: xts## Loading required package: zoo##
## Attaching package: 'zoo'## The following objects are masked from 'package:base':
##
## as.Date, as.Date.numeric##
## Attaching package: 'PerformanceAnalytics'## The following object is masked from 'package:graphics':
##
## legendchart.Correlation(cor(telco[,4:8]), histogram = TRUE)## Warning in par(usr): argument 1 does not name a graphical parameter
## Warning in par(usr): argument 1 does not name a graphical parameter
## Warning in par(usr): argument 1 does not name a graphical parameter
## Warning in par(usr): argument 1 does not name a graphical parameter
## Warning in par(usr): argument 1 does not name a graphical parameter
## Warning in par(usr): argument 1 does not name a graphical parameter
## Warning in par(usr): argument 1 does not name a graphical parameter
## Warning in par(usr): argument 1 does not name a graphical parameter
## Warning in par(usr): argument 1 does not name a graphical parameter
## Warning in par(usr): argument 1 does not name a graphical parameter
Tarea de sesion
- Aplicar transformacion de variables (vistas en clases) para las variables cuantitativas del caso que no se aplico (aplicar para monto, minutos,….)
- Comparar resultados y comentar que transformacion es la que permite mejorar la simetria
- Realizarlo en su cuenta de rpubs y enviar enlace
Fecha maxima de envio: 11/01/2024 – 11:59 pm Modo de envio: Plataforma EVA (solo enlace)