International Journal of Scientific Research and Engineering Development

Employee attrition represents a major challenge for organizations as it directly impacts workforce stability, productivity, and operational efficiency. High attrition rates lead to increased recruitment costs, loss of organizational knowledge, and disruption of business continuity. With the increasing availability of human resource analytics data, machine learning techniques have been widely adopted to predict employee attrition. However, most existing approaches focus primarily on static binary classification models that predict whether an employee will leave or stay, without considering interpretability, temporal risk patterns, or actionable insights. This research proposes a predictive analytics framework for employee attrition that integrates machine learning models with explainable artificial intelligence (XAI) techniques. The frame work utilizes an enhanced HR analytics dataset containing employee demographic, behavioural, performance, and organizational attributes. Random Forest and Logistic Regression models are used as predictive models, while SHAP (SHapley Additive Explanations) is employed to provide global and local interpretability of the predictions. In addition to predictive modelling, this study introduces an Attrition Controllability Index (ACI) to differentiate between controllable organizational factors and uncontrollable demo graphic attributes contributing to attrition risk. Furthermore, the study evaluates fairness and bias across demographic groups to ensure responsible AI deployment in HR decision-making systems. Experimental results demonstrate that explainable machine learning models provide better transparency and actionable insights compared to conventional black-box prediction systems. The proposed framework supports HR managers in identifying key drivers of attrition and implementing targeted employee retention strategies.

Shifa Bilal Tamboli, Simeen Phiroj Mulani, Arman Tajuddin Shiakh,"Predictive Analytics for Employee Attrition Using Explainable AI and Actionable Risk Modelling" International Journal of Scientific Research and Engineering Development, V9(2): Page(2495-2504) Mar-Apr 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.