Prediction of Employee Attrition Using GWO and PSO Optimised Models of C5.0 Used with Association Rules and Analysis of Optimisers

Abstract

Prediction of employee attrition based on five selected attributes which are Gender, Distance from Home, Environment Satisfaction, Work–Life Balance and Education Field out of 36 variables present in the data-set. Application of Grey Wolf Optimisation (GWO) Algorithm and Particle Swarm Optimisation (PSO) on the model of Decision Tree Algorithm “C5.0” which is fed in the inputs of Associated Rules, using this optimised algorithm for the prediction of employee attrition using IBM Watson Human Resource Employee Attrition Data. After comparing the efficiency of GWO and PSO, we have come to a conclusion that time to predict an employee attrition and consumption of RAM have been optimised with GWO. Employee Attrition is one of the major problems faced by companies nowadays. Sometimes, when the long-term working employees leave the company, it affects the relationship of the company with the client and in turn affects the revenue of the company if the person replacing the old employee is not able manage a good rapport with the client. The paper can be used to frame better work policies which will help both the employer and employee. It can be seen as a mirror to the working conditions of the employees.