AI-powered Predictive Modeling for Service Employees: Evidence from an Online Labor Market Platform
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Predicting influential factors in the labor market is a crucial task for companies to obtain and maintain competitive advantages. Nevertheless, little attention has been paid to prediction approaches in understanding influential factors in the labor market. Thus, the purpose of this study is to build prediction models for labor demand, hires, and employee turnover, drawing upon system theory, in the service industry. In specific, the prediction was facilitated by predictors: Economic-related, social-related, labor-related, and industry-related factors. The factors performed prediction for three dependent variables: Labor demand, hires, and turnover. Datasets were collected through the U.S. Bureau of Labor Statistics, Census, Transportation Statistics, Glassdoor.com. Numerous machine learning techniques, including LASSO and random forest regression, performed for better prediction. The findings showed that GLS regression put emphasis on the salary levels in predicting labor demand and turnover. Furthermore, machine learning algorithms and SHapley Additive explanation analysis confirmed the GLS’s results and shed light on other predictors, including age and consumerprice-index. This study provides several implications by featuring the significance of salary levels for increasing hires and decreasing employee turnover. The implications also will help policymakers in developing effective policies to deal with the labor shortage in the industry.
Abstract
Predicting influential factors in the labor market is a crucial task for companies to obtain and maintain competitive advantages. Nevertheless, little attention has been paid to prediction approaches in understanding influential factors in the labor market. Thus, the purpose of this study is to build prediction models for labor demand, hires, and employee turnover, drawing upon system theory, in the service industry. In specific, the prediction was facilitated by predictors: Economic-related, social-related, labor-related, and industry-related factors. The factors performed prediction for three dependent variables: Labor demand, hires, and turnover. Datasets were collected through the U.S. Bureau of Labor Statistics, Census, Transportation Statistics, Glassdoor.com. Numerous machine learning techniques, including LASSO and random forest regression, performed for better prediction. The findings showed that GLS regression put emphasis on the salary levels in predicting labor demand and turnover. Furthermore, machine learning algorithms and SHapley Additive explanation analysis confirmed the GLS’s results and shed light on other predictors, including age and consumerprice-index. This study provides several implications by featuring the significance of salary levels for increasing hires and decreasing employee turnover. The implications also will help policymakers in developing effective policies to deal with the labor shortage in the industry.