Research On Employee Turnover Prediction Method Based On Feature Selection And Machine Learnin

In enterprise management activities,people are the most active and positive constituent elements.The frequent departure of personnel will bring huge losses to the company.According to the report,the employee turnover rate in 2022 is as high as 17.9 %.Especially with the recent unblocking of the epidemic,the problem of employee turnover will be more serious,and how to solve this problem has become a difficult problem for managers.In the era of the digital economy,if enterprises make full use of science and technology to predict employee turnover dynamics and gain insight into employee turnover intention in advance,they can help human resource teams to better respond to talent retention or reserve,which is of far-reaching significance to the development of enterprises.In the field of employee turnover prediction,there have been a lot of research results,but there are still some shortcomings,such as the lack of research methods in feature selection.In view of the shortcomings,this thesis combines the feature selection method based on Improved Genetic Algorithm with the Machine Learning Model to establish an employee turnover prediction model,which better realizes the prediction of employee turnover.This thesis takes the employee turnover data set of Datacastle platform as the research object.Through the preprocessing of data,it mainly includes preliminary data processing,data correlation analysis,feature construction,processing of non-numerical data and numerical data,and processing of unbalanced data.Then,aiming at the problems of "premature convergence" and "slow search" in the Standard Genetic Algorithm,an Improved Genetic Algorithm is proposed,and the Standard Genetic Algorithm and Improved Genetic Algorithm are applied to feature selection,compared with selecting all feature.On this basis,Decision Tree Model,Random Forest Model,XGBoost Model and Light GBM Model are applied to establish the employee turnover prediction model.The experimental results show that the Improved Genetic Algorithm is superior to all features and Standard Genetic Algorithm in calculation accuracy,and superior to Standard Genetic Algorithm in convergence speed,and the Random Forest Model is outstanding under all conditions.Especially the Random Forest Model based on improved Genetic Algorithm,its average accuracy reached 93.58 %.At the same time,in order to give full play to the full performance of the Random Forest Model,this thesis optimizes the hyperparameters of the Random Forest Model with Grid Search technology,improves the average accuracy of the model to 93.86%,and further improves the prediction performance of employee turnover.Finally,the work content is summarized,two suggestions are put forward,and the shortcomings and prospects are pointed out.