Predicting Employee Attrition An Integrated Approach Using Machine Learning and Organizational Factors

This study presents an innovative approach to predicting employee attrition by incorporating temporal dynamics with big data analytics and deep learning techniques. Using the IBM HR Employee Attrition dataset, we developed and evaluated multiple machine learning models, with Random Forest achieving 97.6% accuracy using SMOTE oversampling. The research identifies key temporal predictors including performance trajectories, compensation changes, and work-life balance fluctuations. Our analysis demonstrates how combining temporal dynamics with deep learning can significantly improve predictive capabilities for employee attrition. The study provides both theoretical insights and practical applications for organizations seeking to enhance their workforce management strategies and promote sustainable employment practices. The findings contribute to the broader fields of human resource analytics, machine learning, and organizational sustainability. Keywords: employee attrition, temporal dynamics, deep learning, big data analytics, HR analytics, machine learning, workforce management, LSTM, predictive modeling