Predictive Analytics for Employee Attrition and Strategic Retention Interventions

Abstract

Employee turnover presents significant challenges for many organizations. It reduces productivity, morale, and profitability while driving up recruiting & training costs. For this project, we are going to predict employee retention and understand what factors contribute most to an employee leaving the company by using statistical techniques from historical HR data. We also developed a predictive model using the Random Forest algorithm with under-sampling to consider class imbalance and predict accurately which employees are at risk of leaving the organization. Consequently, these findings may be hints at how they could input interventions around ways through which employee satisfaction and retention levels are increased to reduce turnovers in achieving a sustained workforce-company wealth that will enable them to meet their productivity targets.