Strategic Human Resource Analytics with Explainable Artificial Intelligence: An Interpretable Prediction Framework for Employee Promotion to Support Managerial Decision-Making
Year: 2025
Type: conference-paper
Abstract: Employee promotion decisions profoundly impact organizational performance and employee retention, yet traditional human judgment suffers from subjective biases and inconsistent criteria. This study proposes an explainable artificial intelligence (XAI) framework for strategic promotion prediction, integrating feature engineering, ensemble machine learning (Logistic Regression, SVM, Random Forest, Gradient Boosting), and post-hoc interpretability via SHAP and LIME. Analyzing 54,808 employee records across 11 HR attributes, we demonstrate that Gradient Boosting achieves optimal discrimination (AUC=0.92) and clinical utility (net benefit=0.42), outperforming baseline models by 15%. Calibration analysis reveals systematic overconfidence in Random Forest predictions, necessitating probabilistic recalibration for deployment. SHAP global importance rankings identify training proficiency (|φ|=0.28), prior performance (|φ|=0.21), and awards (|φ|=0.14) as primary promotion drivers, while LIME local explanations provide actionable feedback for candidate development. Our framework advances human resource analytics by reconciling predictive accuracy with decision transparency, enabling evidence-based talent management while mitigating algorithmic bias concerns.
Language: English
Topic: AI and HR Technologies
Domain: Social Sciences
Open Access status: hybrid
OpenAlex ID: W7147635954