Effective machine learning, Meta-heuristic algorithms and multi-criteria decision making to minimizing human resource turnover.

Abstract

Employee turnover is one of the most important issues in human resource management, which is a combination of soft and hard skills. This makes it difficult for managers to make decisions. In order to make better decisions, this article has been devoted to identifying factors affecting employee turnover using feature selection approaches such as Recursive Feature Elimination algorithm and Mutual Information and Meta-heuristic algorithms such as Gray Wolf Optimizer and Genetic Algorithm. The use of Multi-Criteria Decision-Making techniques is one of the other approaches used to identify the factors affecting the employee turnover in this article. Our expert has used the Best-Worst Method to evaluate each of these variables. In order to check the performance of each of the above methods and to identify the most significant factors on employee turnover, the results are used in some machine learning algorithms to check their accuracy in predicting the employee turnover. These three methods have been implemented on the human resources dataset of a company and the results show that the factors identified by the Mutual Information algorithm can show better results in predicting the employee turnover. Also, the results confirm that managers need a support tool to make decisions because the possibility of making mistakes in their decisions is high. This approach can be used as a decision support tool by managers and help managers and organizations to have a correct insight into the departure of their employees and adopt policies to retain and optimize their employees.