Research and Development Employees Attrition: A Machine Learning Based Exploration

Abstract

Technical employees are the most vital asset of a company’s R&D department to remain sustainable, their attrition is a significant concern from the company’s perspective. Despite that technical employee attrition transpires for numerous causes, making it challenging for the predictive high authorities of the R&D department to anticipate these indicators in advance. The attrition of R&D employees affects knowledge retention, project continuity, and the pace of innovation. This research proposes a methodology for predicting attrition among R&D employees, enabling proactive workforce management rather than relying on retrospective metrics. This study developed a predictive Machine learning classification model utilizing 30 features that influence R&D employees attrition, derived from the IBM dataset, comprising 961 records. Consequently, five Machine learning classification predictive models- Decision Tree, Random Forest, Support Vector Machine, Logistic Regression, and K-Nearest Neighbour were developed and their performance assessed. Moreover, the analysis of variables affecting R&D employees’ attrition revealed that stock option level, job involvement, job level, and job satisfaction were the most significant contributors. This research helps the organizations and managers to identify the attrition factors by using the machine learning models and develop sustainable strategies on retention.