Analysis and Prediction of Employee Promotions Using Machine Learning

Predicting employee performance is essential for organizations. The success or failure of a company often depends upon the competence of its employees, so CEOs and managers who want their organizations to succeed face the difficult task of determining which employees should be promoted. The current promotion process used in most organizations should be considered misleading because it depends on supervisors' judgments. The major aim of this paper is to use classification algorithms to develop predictive models for predicting whether an employee is qualified for promotion or not and identifying the most important attributes affecting employee promotion. The dataset used in this paper is from Kaggle 2020. It contains information on multinational companies arranged in 54,808 rows and 13 columns. This dataset covers nine broad verticals across organizations. Several predictive modeling techniques, including K-Nearest Neighbors, Logistic Regression, Decision Tree, Random Forest, Support Vector Machine, and Ensemble models (Adaboosting and Gradient Boosting models) were used to predict employee promotion. Based on measurements of accuracy, Fl-score, and AUC, Gradient Boosting outperforms the other classification algorithms. The results also show that the most significant factor contributing to predicting employee promotion is the previous year's rating. Department had no effect on employee promotions.