Binomial Logistic Regression and XGBoost Model of Multiple Factors on Employee Well-being

Employee Well-being is the physical and psychological experience and feeling of employees during work, it is a critical indicator of employee's quality of life and plays an important role in organization development. Therefore, it is of great significance to explore the main factors that affect Employee Well-being and learn the mechanism between Employee Well-being and its influences. Binomial Logistic Regression and XGBoost are machine learning tools using different algorithm. Binomial Logistic Regression uses linear regression analysis to integrate linear combination of features with logical functions, while XGBoost is a gradient Boosting algorithm, it uses decision trees as base learners and integrates their predictions to form powerful integrated models. In this study, Binomial Logistic Regression model was built to get a multivariate linear relationship between dependent variable and independent variables and XGBoost model was set to get a multivariate nonlinear relationship between dependent variable and independent variables. 6 independent variables, Organizational Innovation, Cooperative Behaviors, Job Satisfaction, Sense of Control, Perceived Stress and Trust which showed significant correlations with Employee Well-being were obtained by Pearson correlation analysis. Both Binomial Logistic Regression model and XGBoost model simulated well in accuracy, precision, recall and F1score, and demonstrated explicit classification and prediction effect. Perceived Stress showed the most significance to Employee Well-being in both models, which meant the main factor influencing dependent variable. Since the correlation between Employee Well-being and its main factors were explored successfully by the two models, it would be a simplified and efficient way to apply machine learning model to explore the mechanism of Employee Well-being with its potential factors, predict the level of Employee Well-being accurately, adjust the status of Employee Well-being timely.