A Learning Material-Based Recommendation System For E-Learning

Abstract

There may be many kinds of e-learning material presented to learners. However, learners may not have the consciousness or not want to spend time to examine to select the most appropriate one. Recommendation systems seem to be a feasible solution for an efficient learning process both for learners and the service provider. In this study, we propose a learning material-based e-learning recommendation system that considers the learners' learning material preferences and uses the collaborative filtering method for the recommendation system. To obtain realistic results, actual data gained from Anadolum eCampus, the learning management system of Anadolu University Open Education System, were used. In addition, this study aimed to select the most successful algorithm by applying three Collaborative Filtering (CF) algorithms (kNN, k-means and SVD-based CF) in the experiments to keep the efficiency high. As a result, k-means and SVD-based CF algorithms were more successful than kNN-based CF algorithms. In addition, the SVD-based CF algorithm was the most successful regarding speed performance. In conclusion, this system can be used in e-learning settings to recommend learning materials to learners according to their preferences.