Comparative Prediction of Student Course Preference Using DT-KNN Hybrid Model

This study aims to predict students' course preference levels (categorized into three levels: high, medium, and low) using Decision Tree (DT) and K-Nearest Neighbors (KNN) algorithms by analyzing academic performance, attendance rates, classroom participation, and regular assignment data. Based on the analytical results, targeted teaching optimization strategies will be proposed. Distinct datasets were employed to train DT and KNN classification models respectively, with systematic evaluation and comparison of their predictive performances. The research demonstrates that the DT algorithm exhibits superior practical value in course preference prediction, with the experimental model achieving an accuracy rate exceeding 90%.