Computer modeling of company employee churn using machine learning and predictive analytics methods

Abstract

The study proposes an approach to management employee churn using machine learning, predictive analytics methods and with the support of information technologies. It systematically explores the theoretical underpinnings of employee turnover, categorizing it into various types and identifying key components for measurement, as well as factors influencing churn rates and their potential regulation. The discourse highlights the importance of employee retention strategies, emphasizing elements such as compensation, work planning, performance evaluations, training programs, and opportunities for career advancement. The authors propose utilizing decision tree and logistic regression methodologies to predict employee churn, selecting a binary classification criterion that distinguishes between employees who remain with the company and those who depart. Two predictive models are developed, showcasing significant accuracy metrics: the decision tree model demonstrates an impressive 91.3% accuracy on training data and 74.19% on test data, while the logistic regression model indicates 88.41% accuracy on training data and 90.32% on test data. These findings underscore the reliability of the proposed models in forecasting employee turnover. Furthermore, the article outlines practical applications for these models, providing actionable insights for organizations aiming to implement data-driven strategies for improving employee retention. By leveraging advanced analytical techniques, the study contributes valuable methodologies and frameworks for companies grappling with the challenges of employee churn, positioning itself as a critical resource for human resource professionals and organizational leaders seeking to enhance workforce stability and performance. Overall, the research underscores the potential of predictive analytics in informing strategic decisions regarding employee retention and maintaining a healthy climate in the company.