Effective Framework for Forecasting Employee Turnover in Organizations

Abstract

Employee turnover is a crucial issue for businesses, producing significant expenses and problems. Accurate turnover prediction is vital for building talent retention strategies. Conventional models frequently struggle to account for the changing and nonlinear elements that affect employee turnover. In this study, a novel Dynamic Black Hole tuned Logistic Regression (DBH-LR) was proposed to accurately forecast employee turnover in the organization. The employee turnover dataset was gathered from the Kaggle source. Data cleaning is employed to ensure the quality and dependability of the employee dataset. The Black Hole conditions in space, which dynamically modify the locations of potential solutions to converge towards an ideal solution, served as the model for the DBH algorithm. The Logistic Regression (LR) model's parameters are adjusted using DBH to increase the model's forecasting accuracy. The suggested DBH-LR method is tested on the Pythonplatform. The DBH-LR model outperformed existing techniques in terms of recall (97.4%), accuracy (96.3%), F1-score (95.3%), and precision (96.1%). Future studies may examine the use of this methodology in different organizations and worker groups to validate its efficiency.