Carga de datos
datos <- read.csv("C:/Users/sarit/Downloads/WA_Fn-UseC_-HR-Employee-Attrition.csv", header=TRUE)
Modelo 1 - En origen
modelo1 <- lm(MonthlyIncome ~ Age + TotalWorkingYears + YearsAtCompany + JobLevel + DistanceFromHome + Education, data=datos)
summary(modelo1)
##
## Call:
## lm(formula = MonthlyIncome ~ Age + TotalWorkingYears + YearsAtCompany +
## JobLevel + DistanceFromHome + Education, data = datos)
##
## Residuals:
## Min 1Q Median 3Q Max
## -5501.4 -927.6 72.3 762.4 4029.0
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -1378.939 203.337 -6.782 1.72e-11 ***
## Age -9.331 5.816 -1.604 0.10883
## TotalWorkingYears 61.062 10.161 6.009 2.35e-09 ***
## YearsAtCompany -15.209 8.105 -1.877 0.06078 .
## JobLevel 3792.925 54.925 69.057 < 2e-16 ***
## DistanceFromHome -12.746 4.652 -2.740 0.00622 **
## Education -23.001 37.651 -0.611 0.54136
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1445 on 1463 degrees of freedom
## Multiple R-squared: 0.9062, Adjusted R-squared: 0.9058
## F-statistic: 2355 on 6 and 1463 DF, p-value: < 2.2e-16
Modelo 2 - Semi-logarítmico
modelo2 <- lm(log(MonthlyIncome) ~ Age + TotalWorkingYears + YearsAtCompany + JobLevel + DistanceFromHome + Education, data=datos)
summary(modelo2)
##
## Call:
## lm(formula = log(MonthlyIncome) ~ Age + TotalWorkingYears + YearsAtCompany +
## JobLevel + DistanceFromHome + Education, data = datos)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.99047 -0.16108 0.00605 0.16951 0.73597
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 7.336e+00 3.648e-02 201.076 <2e-16 ***
## Age 1.242e-03 1.043e-03 1.190 0.2341
## TotalWorkingYears 2.665e-03 1.823e-03 1.462 0.1440
## YearsAtCompany 1.275e-03 1.454e-03 0.877 0.3806
## JobLevel 5.272e-01 9.855e-03 53.496 <2e-16 ***
## DistanceFromHome 2.603e-05 8.347e-04 0.031 0.9751
## Education 1.482e-02 6.756e-03 2.194 0.0284 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.2593 on 1463 degrees of freedom
## Multiple R-squared: 0.8484, Adjusted R-squared: 0.8477
## F-statistic: 1364 on 6 and 1463 DF, p-value: < 2.2e-16
Modelo 3 - Log-log
modelo3 <- lm(log(MonthlyIncome) ~ log(Age) + log(TotalWorkingYears+1) + log(YearsAtCompany+1) + JobLevel + DistanceFromHome + Education, data=datos)
summary(modelo3)
##
## Call:
## lm(formula = log(MonthlyIncome) ~ log(Age) + log(TotalWorkingYears +
## 1) + log(YearsAtCompany + 1) + JobLevel + DistanceFromHome +
## Education, data = datos)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.98905 -0.16355 0.01392 0.15706 0.72792
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 7.4177745 0.1129634 65.665 <2e-16 ***
## log(Age) -0.0811648 0.0366332 -2.216 0.0269 *
## log(TotalWorkingYears + 1) 0.1695769 0.0178771 9.486 <2e-16 ***
## log(YearsAtCompany + 1) 0.0074854 0.0111632 0.671 0.5026
## JobLevel 0.4828471 0.0082755 58.347 <2e-16 ***
## DistanceFromHome -0.0001992 0.0007991 -0.249 0.8031
## Education 0.0095436 0.0065050 1.467 0.1426
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.2481 on 1463 degrees of freedom
## Multiple R-squared: 0.8611, Adjusted R-squared: 0.8605
## F-statistic: 1512 on 6 and 1463 DF, p-value: < 2.2e-16
Normalidad de los residuos
shapiro.test(residuals(modelo1))
##
## Shapiro-Wilk normality test
##
## data: residuals(modelo1)
## W = 0.99159, p-value = 1.904e-07
Multicolinealidad
library(car)
vif(modelo1)
## Age TotalWorkingYears YearsAtCompany JobLevel
## 1.985684 4.397573 1.734587 2.600508
## DistanceFromHome Education
## 1.000599 1.046103
Heterocedasticidad
library(lmtest)
bptest(modelo1)
##
## studentized Breusch-Pagan test
##
## data: modelo1
## BP = 84.069, df = 6, p-value = 5.146e-16
Corrección con errores robustos
library(sandwich)
coeftest(modelo1, vcov = vcovHC(modelo1, type = "HC1"))
##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -1378.9387 189.2262 -7.2873 5.155e-13 ***
## Age -9.3307 5.3997 -1.7280 0.084201 .
## TotalWorkingYears 61.0623 11.9708 5.1009 3.822e-07 ***
## YearsAtCompany -15.2090 9.6917 -1.5693 0.116798
## JobLevel 3792.9249 57.7151 65.7180 < 2.2e-16 ***
## DistanceFromHome -12.7458 4.7175 -2.7018 0.006976 **
## Education -23.0011 38.9400 -0.5907 0.554825
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1