Machine Learning-Based Recommendation System for E-Learning: A Personalized Approach Using Learner Behavior Analytics

Abstract

 Modern learning is rapidly evolving and has transformed the way we learn today, offering flexibility and scalability. Current recommendation systems are typically based on "one size fits all" approach and do not consider individual learners, their preference and varying levels of understanding. Lack of personalization in e-learning systems can result in low engagement, inefficient learning, and low performance. To overcome this problem, the current article introduces a recommendation system based on machine learning (ML) that is used to provide individualized learning materials based on the behavior of the learners. The suggested model takes advantage of the data regarding learner interaction, such as browsing history, assessment results, and content interests, and constructs an adaptive recommendation system. It uses a hybrid strategy which integrates collaborative filtering and content-based strategies to achieve better accuracy of the recommendations and to overcome the constraints like cold start problem. Moreover, the system is effective in changing dynamically to learner profiles, resulting in better engagement and knowledge retention. The main contribution of the current article is the integration of behavioral analytics with hybrid ML models that are used to improve the personalization of the e-learning environment. The proposed hybrid ML recommendation system achieved superior performance with 88% accuracy, 0.85 precision, 0.83 recall, and lowest RMSE (0.35). Results indicate that the given system could be implemented to facilitate adaptive learning and provide overall better learning outcomes.