A Detailed Analysis of AI Models for Predicting Employee Attrition Risk

Employee attrition is one of the biggest problems of all organizations in today's world. Typical organizations have 12-15% attrition on average. Replacing an employee is very expensive and it puts stress on the teams by impacting the morale and also resulting in unnecessary overtime for them. The average hiring cost of a software engineer is approximately $40000. In light of these facts, it is clear that organizations need to find ways to control or reduce attrition. A first step to controlling attrition lies in predicting the attrition risk of the employees. In this paper, we analyze the prominent factors impacting employee attrition using the IBM HR Analytics Data set from Kaggle with various machine learning models to predict the attrition risk. We also compare the accuracy of these models with respect to the Area Under Curve (AUC). We select the main factors affecting employee attrition by using Random Forest, and classify which types of people are more likely to quit by utilizing the Xg boost classification. We also discuss the approaches that an organization can use to keep its employees engaged.