Abstract
Negative academic emotions reflect the negative experiences that learners encounter during the learning process. This study aims to explore the effectiveness of machine learning algorithms in predicting high school students' negative academic emotions and analyze the factors influencing these emotions, providing valuable insights for promoting the psychological health of high school students. Based on the microsystem proposed in ecological systems theory, we comprehensively consider individual and school factors that affect students' negative academic emotions. We randomly selected 1,710 high school students from Hebei Province, China (742 males), who completed the Adolescent Resilience Scale, Multidimensional Multi-Attributional Causality Style Scale, Academic Self-Efficacy Questionnaire, Teacher Discipline Style Scale, and Academic Emotion Scale. We applied various machine learning models, such as logistic regression, naive Bayes, support vector machine, decision tree, random forest, gradient boosting decision tree, and adaptive boosting, to analyze the students' negative academic emotions. The results show that the random forest model had the best predictive performance, with an accuracy of 83.9%. Subsequently, the importance of variables was determined using the forward feature selection method. We concluded that the most important factors for predicting high school students' negative academic emotions are affect control, followed by ability attribution, luck attribution, background attribution, self-efficacy for learning behaviors, and self-efficacy for learning abilities. This study validates the applicability and value of machine learning models in predicting negative academic emotions, providing important insights for educational practice. When designing intervention strategies, attention should be given to the development of emotional control and attribution styles to help students better alleviate excessive negative academic emotions.