Predictive Modelling of Employee Attrition using HR Analytics

Ms. R. Javi Prabha MBA

NET, Assistant Professor Dhanalakshmi Srinivasan University, Samayapuram

Sudharsana S S

MBA – Business Analytics and Human Resource

ABSTRACT:Employee attrition is a persistent and costly challenge in the Information Technology sector. This article presents a data-driven study at Extazee Software Solution Company to predict and manage employee turnover using HR analytics and machine learning. Based on structured survey responses from 100 employees, the study analyses key attrition-influencing variables—including job satisfaction, salary adequacy, and work environment quality—through Chi-Square inferential testing and three classification algorithms: Logistic Regression, Decision Tree, and Random Forest.Statistical analysis confirmed significant associations between job satisfaction (χ² = 17.81, p < 0.001) and work environment quality (χ² = 9.90, p = 0.007) with attrition, while salary showed a directional but non-significant trend (χ² = 5.23, p = 0.073) at the 5% level. Feature importance analysis identified job satisfaction, active job-search behaviour, and work-life balance as the strongest attrition predictors.With 49.1% of employees having considered leaving in the previous six months and 61.4% conditionally open to departure, the findings carry urgent operational significance. The article translates these results into actionable HR retention strategies and argues that predictive HR analytics is accessible and impactful even in small IT organisations.

Keywords: employee attrition, HR analytics, machine learning, predictive modelling, job satisfaction, Chi-Square test, Random Forest, workforce retention, IT sector