in our data we have information about the people who had left the company and about the people who are still working. We will use this information to fit a logistic regression model to predict the probablity of attrition for a given employee given the related information associated with the employee. The dependent variable for the model is of course "attrition", which is a binary variable with "yes" and "no" outcome.

We will begin with all the variables in the model and then use the forward or backward selection method to select a good model.

R Code:

fullmod <- glm(Attrition ~ .,family = binomial,data=HR)

summary(fullmod)

Output: