A Novel Fuzzy Decision-Making Model for Employee Attrition Prediction using Ensemble Learning Techniques

Employee attrition is a significant challenge for organizations, affecting productivity, knowledge retention, and recruiting costs. This research contains a hybrid framework utilizing Random Forest (RF), Forest Optimization Algorithm (FOA), and fuzzy decision making to accurately predict employee attrition and other talent management functions. FOA helps enhance the accuracy of RF by optimizing the feature selection, experience, and tuning of the RF hyperparameters, while fuzzy rules address the uncertainties surrounding HR attributes such as job satisfaction, overtime, and work-life balance, providing outputs that managers can interpret and utilize to make decisions. Experiments using the IBM HR Analytics dataset demonstrate the proposed model achieves 93.4% accurate and an AUC of 0.92, producing estimates that outperform flexible baseline models including Logistic Regression, Decision Tree, and standard Random Forest. The results outlined in this research help demonstrate the benefits of optimization and fuzzy logic to achieve highly accurate and interpretable predictors, recommending explicit actions to allow for proactive employee retention.

Hali tarjima qilinmagan