Study on Online Learning Behavior Analysis and Performance Prediction Based on Improved Random Forest Algorithm

With the popularity of online learning, analysis student learning behavior and prediction academic performance have become a hot research topic in field education. This study aims to analyze factors influencing outcomes on platforms, extract feature variables from behavior, construct weighted random forest model predict students’ performance. Additionally, method for computing importance is designed based significance different behavioral features. contributes gaining early insights into progress, proactively identifying issues strategically allocating educational resources, thus providing students with enhanced guidance support.