RECOMMENDATION SYSTEM FOR PERSONALIZED LESSON LEARNING

Abstract

 This research presented the development of recommendation system for a personalized lesson recommendation system in online learning environments. The system combined ontology techniques to represent learner knowledge and lesson content with machine learning techniques to predict suitable lessons based on individual behaviors and characteristics. The developed system used information from students in the information technology course through an e-Learning platform and evaluated performance by comparing it with traditional Content-Based Filtering and Collaborative Filtering techniques. The results showed that the system could increase the average F1-score to 0.80 and reduce Mean Absolute Error (MAE) by an average of 18 percent compared to traditional methods. Additionally, students showed significantly improvement in academic achievement (p < 0.05) and reported high satisfaction with the system. This research demonstrated that combining ontology with machine learning algorithms enhanced the accuracy of lesson recommendation and helped promote effectively personalized learning.