An efficient Content-Based e-Learning Recommender System using K-L Divergence and Sentiment Score
Abstract
E-learning is of paramount importance these days. The enormous growth of e-learning content has resulted into the rise of recommender systems.E-learning recommender systems are required for recommending the best recommendations according to the need of an individual.Raj Kumar and Bhatia have calculated cosine similarity for generating recommendations for a particular student [1]. The cosine similarity is used to know the directional similarity between the courses. However, keeping in mind the probabilistic nature of data under consideration. It will be better to use K-L divergence for more accurate recommendations. In this paper, ordinal data has been used and K-L divergence is applied to achieve more accurate recommendations. Further, the sentiment score associated with course is applied to improve efficiency of the course recommendations.