College Ideological and Political Education System based on Improved Random Forest Model and Content-based Filtering

In recent years, the rapid growth of Information Technology (IT) has transformed education sector particularly College Ideological and Political Education System (CIPES) to enhance teaching quality and engaging learning experience. However, education sector faces many challenges and complexities in CIPES, lacking quality learning and sufficient career support which inhibits improved educational outcomes. To overcome these challenges, this research proposed Improved Random Forest (IRF) model to predict student interest in CIPES courses and develop accessible platform for learning and teaching. Initially, the data is collected from EdX dataset that comprises records of 976 courses, then preprocessed with stop-words removal and lemmatization technique by removing unwanted data. Next, Term Frequency-Inverse Document Frequency (TF -IDF) method is used to extract the features based on term frequency. After that, IRF model utilized Recursive Feature Elimination (RFE) method to select most important features by eliminating less important features to forecast student interest in CIPES courses. It improves system ability to generalize by reducing data overfitting, then Content-Based Filtering (CBF) is designed to provide personalized course recommendations based on predictions. Form the results, the proposed IRF model achieved outstanding results in terms of accuracy (92.21 %) and recall (93.63%) when compared with the existing Primary and Secondary Education Resource System by using Collaborative Filtering (PSERS-CF) model.