Employee Churn Estimation Using Machine Learning Methods

Abstract

Employee churn is a vital concern for organizations. Many machine learning (ML) based systems have been improved to bring solution to the employee churn problem. The present study tries to predict employee churn using ML algorithms. From this point on, the aim of the study is to predict employee churn for organizations based on mathematical methods using machine learning. Application of information technologies in Human Resources Management in organizations, the data collected is considered as a critical factor in making decisions about employees. The sample of the study consists of 49,653 employees from different sectors. There are almost no studies using machine learning algorithms in management studies. In this sense, it is predicted that the study will fill an important gap in terms of contributing to the literature. In this paper, logistic regression, random forests, artificial neural networks and support vector machine methods were used as classification methods in the research to estimate whether an employee will quit or not. The results are compared with other Machine Learning algorithms. In the analysis, it was seen that the random forests method was found to be the most reliable machine learning with a rate of 98.7715 % in estimating the employee quitting.