Ensembles of decision trees and gradient-based learning for employee turnover rate prediction

Abstract

Employee turnover has a negative impact on business profitability. To tackle this issue, we can utilize computational advancements to forecast attrition and minimize expenses. We employed an HR Analytics dataset to investigate the feasibility of using these predictive models in decision support systems. We developed an ensemble of gradient-based decision trees that accurately predicted employee turnover and performed better than other sophisticated techniques. This approach demonstrates exceptional performance in handling structured and imbalanced data, effectively capturing intricate patterns. Gradient-based decision trees provide scalable solutions that effectively balance predictive accuracy and computational efficiency, making them well-suited for strategic business analysis. The importance of our findings lies in their ability to offer dependable insights for making well-informed decisions in business settings.

Keywords

• Decision trees
• Gradient learning
• Data science
• Turnover rate prediction
• Machine learning