Predicting Employee Attrition using Random Forest

Abstract

This paper aims to address the issue of employee retention in today's businesses. However, they are unable to identify the actual causes of their resignation. That could be due to a variety of factors, such as: financial, cultural, etc.). How a company treats its employees and ensures their happiness is unique to each business. However, the satisfaction rate is frequently ignored. As a result, employees frequently quit their jobs abruptly and without giving a reason. Researchers have become increasingly interested in machine learning (ML) methods over the past few decades. It is capable of suggesting solutions to numerous issues. Then, machine learning has the potential to anticipate employee attrition by making predictions. Using a real data set with a sample size of 1469, the authors of this paper compare cutting-edge options for the proposed machine learning algorithms. Managers could use the findings as a warning to alter their strategies or behavior. It could also be used to suggest new policies to managers in order to keep employees employed by the company. The goal of this study is to compare and contrast various machine learning approaches for predicting which employees will likely leave their company. With nearly 50 useful information units, the data set contains information about both current employees and employees who had already left their jobs. Numerous factors are combined in this last: cultural, professional, social, financial, and relational factors In this paper, six distinct ML algorithms were utilized. The Random Forest algorithm performed the best in terms of predicting employee attrition, as demonstrated by the experiments. The best prediction accuracy, 85.12, is regarded as satisfactory accuracy