Prediction of Employee Loyalty Classification for Maintaining Company Sustainability Using Combination of SEM-PLS and Machine Learning

The company's sustainability is influenced by several internal and external factors, one of them is the turnover intention which has an impact on the weak loyalty of human resources. Based on the results of a survey, it shows the level of turnover intention is influenced by many factors. A high level of turnover intention will threaten the sustainability of the company. This should be of particular concern to human resource managers. Therefore, this study aims to analyze the prediction of turnover intention based on the factors affecting employees. The method used is a combination of SEM-PLS and Machine Learning based on the Random Forest algorithm. The empirical model developed in this study was based on four research variables. The variables used are Job Satisfaction (JS), Organizational Commitment (OC), Employee Performance (EP), Turnover Intention (TI). Then, analysis of the variable relationship using SEM-PLS resulted in JS and OC having a significant effect on the level of turnover intention. Thus, each indicator from JS and OC can be used to create a predictive model of turnover intention classification based on the Random Forest algorithm . The results of the Random Forest model obtained have a very good prediction accuracy of 98%. Based on the analysis of the important variables, it was found that OC01 was the most influential variable on the level of turnover intention, with a mean decrease accuracy of 39.51. The predictor variable OC01 means that employees have a good level of 'strong belief in a career in the company'. Therefore, human resource managers must pay better attention to career projections and the suitability of employee positions. So, the employees will give loyalty to the company, then it can better maintain the company's sustainability