A multifactor model for predicting and managing employee turnover risks at nuclear enterprises

This study examines the process of predicting and managing the risks of voluntary employee turnover at nuclear enterprises. The research focuses on the human resources potential of a high-tech nuclear enterprise. Particular attention is paid to identifying and quantifying risk factors specific to the nuclear industry, such as career path clarity, participation in innovative projects, and working under sensitive conditions, with the development of an interpretable mathematical model. Highly qualified personnel are a key asset for nuclear enterprises, where employee turnover can pose critical risks, including the loss of unique production competencies, reduced safety, and increased operating costs. The objective of the study is to develop a multifactorial model for predicting the individual risk of voluntary employee turnover at nuclear enterprises. The research methodology is based on the analysis of HR management system data. Logistic regression, which ensures highly interpretable results, was used to construct the predictive model. The model was chosen due to its advantages for binary classification problems, such as high interpretability of results, robustness to multicollinearity, and relatively low computational complexity. The study demonstrated that decreased career trajectory clarity increases the risk of dismissal, as does a lack of participation in innovative projects. The developed model allows for segmenting personnel into three risk groups and developing targeted interventions for each category of employee. The key findings include prioritizing organizational, psychological, and career factors over financial incentives. The author's particular contribution lies in the creation of a comprehensive proactive risk management system, including a mechanism for assessing the cost-effectiveness of measures taken. The novelty of the study lies in adapting the survival analysis methodology to the nuclear industry and integrating quantitative forecasts with practical HR strategies. The developed model is an effective tool for HR departments at nuclear enterprises, facilitating the transition from reactive to predictive management of personnel stability.