MACHINE LEARNING-BASED ANALYSIS OF STUDENT PERFORMANCE IN E-LEARNING
Keywords:
Machine Learning, E-Learning, Student Performance Analysis, Learning Analytics, Predictive Modeling, Educational Data Mining, Academic AchievementAbstract
Machine learning allows for the monitoring of online student performance and the identification of learning difficulties. Student activity records, assignment submissions, assessment scores, and login frequency predict learning. Data preprocessing decreases missing values, noise, and class imbalance, enhancing model dependability. Random Forest, Logistic Regression, Decision Tree, and KNN are compared. Academic progress, learning consistency, and engagement can be tracked via feature extraction. The method suggests early identification of at-risk children for academic assistance. Experiments show ensemble models predict better than individual classifiers. We evaluate the model using confusion matrix analysis, F1-score, recall, accuracy, and precision. Examining major predictors enhances interpretation.
