In this analysis employee attrition has been modelled through logistic regression using a range of input variables such as survey data, demographic, working hours and other HR information.
Just like linear regression, logistic regression is the weakest of the classification algorithm but it's a good starting point as it's very easy to explain (its ease of interpretation rivalled only by decision trees) and serves as a baseline performance benchmark. Another hidden advantage (or disadvantage?) of logistic regression is that it predicts probabilities by default and instead of classes, which can be interpreted as required by the user. Step wise variable reduction during logistic regression helps in understanding variable interdependecy and variable importance.