Research on Student Classroom Quality Evaluation Model Based on Decision Tree Expression Recognition and Factor Analysis

This study proposes a novel student classroom performance evaluation model integrating decision tree-based facial expression recognition and factor analysis to address limitations of traditional subjective and non-real-time methods. The model is based on seven classroom behavior indicators derived from 32 students. The C4.5 decision tree algorithm classified 1,500 facial feature datasets with a recognition accuracy of 92.3%. Factor analysis reduced the indicators to three principal components: focus (45.8% variance contribution), activity level (34.6%), and fatigue level (14.8%), with a cumulative variance contribution of 95.3%. A weighted scoring model was developed, providing real-time feedback on student performance. For example, student ID 13 scored 3.983, significantly higher than the class average of 1.521. Focus was found to be the most correlated with classroom quality (r = 0.87). This data-driven tool enables teachers to offer personalized suggestions, such as adjusting teaching pace for fatigued students, thereby enhancing teaching efficiency.