Management and prediction of employee turnover in enterprises based on big data analytics and machine learning

Abstract

In the context of aging population and social transformation, employees also put forward higher demands on corporate culture and values, which also increases the risk of employee turnover. This paper explores the correlation between various indicators and leavers through statistical analysis and visualization of various data indicators and the distribution of leavers and non-leavers on various indicators by using whether they left the company as a classification criterion. In addition, this paper also carries out Pearson correlation analysis for each indicator and draws a correlation heat map to quantitatively explore the correlation between indicators. In order to predict whether employees will leave the company, this paper uses random forest model, support vector machine model, KNN model, plain Bayesian model and logistic regression model for training and testing. The results show that the best prediction in terms of employee turnover is the Random Forest model with a prediction accuracy of 98.8%. This was followed by the Support Vector Machine model with an accuracy of 95.1%. In addition, the KNN model also achieved an accuracy of 94.8%. Ordinary Bayesian model and logistic regression model have lower accuracy rates of 80.4% and 77.2% only. This is of great significance for enterprises to realize sustainable development, and is worthy of in-depth study and practice by enterprise managers.