A New Perspective to Evaluate Machine Learning Algorithms for Predicting Employee Performance

Abstract

Performance prediction is the forecast of future performance conditions based on past and present information. Forecasts can be made about companies, departments, systems, processes, and employees. This study focuses on assessing employee performance in terms of employee behavior, work, and growth potential. Organizations benefit when their employees perform well. Therefore, predicting employee performance plays an important role in a growing organization. To this end, we propose three machine learning algorithms: a support vector machine, a decision tree (j48), and a naive Bayes classifier. These can predict employee behavior in the workplace. Comparing the results, the Naive Bayes algorithm shows better results than the other two algorithms on the basis of metrics such as timeliness, error loss, and accuracy.