Exploration of Predictive Model and Influential Variables for Teachers’Job Satisfaction using Automated Machine Learning

The purpose of this study is to explore the predictive model and influential variables for teachers’ job satisfaction using first-year data from the Seoul Education Longitudinal Study of Teachers (SELST) 2020. For this purpose, 242 explanatory variables, processed through a preprocessing stage, were used for the analysis. The key findings are as follows. First, when applying Automated Machine Learning (AutoML) techniques, the light gradient boosting machine showed the best performance in predicting teachers' job satisfaction. Second, when the key influential variables of teachers' job satisfaction were derived based on the SHAP index, an explainable AI technique, the top 20 explanatory variables were found to be included in the areas of personal characteristics, teacher quality, school characteristics, and policies and systems. Among these, retirement expectation status, creativity, participation in school management, the current school level (elementary school status), and implementation of teacher policies and systems (normalization of school work) were newly identified as the key influential variables of teachers' job satisfaction through this study. Finally, based on the study findings, multifaceted implications for enhancing teachers' overall job satisfaction and suggestions for future studies were provided.