Finding optimal pedagogical content in an adaptive e-learning platform using a new recommendation approach and reinforcement learning

• Published in: Journal of ambient intelligence and humanized computing 2020-10, Vol.11 (10), p.3921-3936
• Main Authors: Madani, Youness, Ezzikouri, Hanane, Erritali, Mohammed, Hssina, Badr
• Format: Article
• Language: English
• Subjects: Artificial Intelligence; Clustering; Collaboration; Computational Intelligence; Data mining; Distance learning; Engineering; Filtration; Genetic algorithms; Literature reviews; Machine learning; Objectives; Online instruction; Original Research; Pedagogy; Platforms; Recommender systems; Robotics and Automation; Sentiment analysis; Similarity measures; Skills; Social networks; Sparsity; User Interfaces and Human Computer Interaction
• ISSN: 1868-5137; 1868-5145
• DOI: 10.1007/s12652-019-01627-1

In the learning process, learners have different skills and each one has his own knowledge and his own ability to learn. The adaptive e-learning platforms try to find optimal courses for learners based on their knowledge and skills. Learning online using e-learning platforms becomes indispensable in the teaching process. Companies and scientific researchers try to find new optimal methods and approaches that can improve education online. In this paper, we propose a new recommendation approach for recommending relevant courses to learners. The proposed method is based on social filtering(using the notions of sentiment analysis) and collaborative filtering for defining the best way in which the learner must learn, and recommend courses that better much the learner’s profile and social content. Our work consists also in proposing a new reinforcement learning approach which helps a learner to find the optimal learning path that can improve the quality of learning.