Modeling Student Learned Helplessness in Mathematics Using Random Forest

This study aims to model and detect student learned helplessness (LH) in Mathematics using random forest. 113 8thgrade students participated in using a web application for step-bystep solving linear equations. The Random Forest model demonstrated outstanding performance. It achieved 92% accuracy and 93 % mean F1 score. Key predictors of LH included GWA, anxiety, and self-efficacy. These three accounted for more than 67 % of the model's predictive power. The model also excelled in balancing precision and recall, with a 98 % recall rate for high LH cases which ensures that most at-risk students were accurately identified. These results revealed the reliability of the model and its precision in detecting LH. This makes it well-suited for a mathematics tutoring system that enables early detection of LH which makes it intelligent and adaptive to its interventions which could lead to improved student success in mathematics.