Machine Learning in Human Resource Analytics: Promotion Classification using Data Balancing Techniques

Nowadays, Machine Learning (ML) is widely used in many business fields, and numerous companies use it to enhance their Human Resource Management Systems (HRMS) systems. By incorporating HR analytics and ML techniques, companies can harness the power of ML throughout the employee life cycle – from recruitment to retirement. This study provides an overview of how ML applications are used in HRMS. Emphasizes the importance of addressing imbalanced data to improve system performance. To tackle this issue, Synthetic Minority Over Sampling Technique (SMOTE) and normalization techniques were employed in this research for promotion classification. The data used for analysis was derived from HR sources on Kaggle (Kaggle.com/datasets). The results obtained from this study were compared with a study that utilized Decision Tree (DT) and Principal Component Analysis (PCA) for promotion classification. The findings reveal that employing the Synthetic Minority Over Sampling Technique (SMOTE) and the Random Forest Classifier (RFC) yields results achieving an accuracy rate of 99%. Exploring models and applying them to diverse datasets would be beneficial for researchers. This study comprehensively reviews existing evidence on machine learning within HR analytics. It highlights how data processing techniques, like SMOTE and PCA, can effectively address imbalanced data while delivering accuracy.