Examining and Predicting Teacher Professional Development by Machine Learning Methods

Abstract

Core quality and ability development analysis is important for evaluating and deepening the professional development of teachers. It is also useful for improving the quality of education and promotes the growth of the next generation. This research tries to investigate how prominent teachers regulate their professional development. The research also attempts to provide suggestions about current professional status of teachers. The answer to these questions will help teachers to better comprehend how teachers develop. To perform factor analysis for teacher professional development, we develop a set of questionnaire scheme, and collected a certain number of samples through the survey. Professional development is transformed to a classification problem. This study applies machine learning (ML) methods to identify significant attributes that prominent teachers shown in class education, and predict the level of teacher professional development. Eight ML methods are taken to classify the samples. Hyperparameter optimization is performed to improve prediction accuracy. The simulation results show that the ensemble method, support vector machine and artificial neural network are the top three ML methods for the problem. Hyperparameter optimization does not show great impact on the performance of the ensemble and support vector machine methods. The accuracy can reach above 85% by tuning the artificial neural network method through hyperparameter optimization. This study provides an important basis for the future intelligent analysis of teacher professional development.