HYBRID EDUCATIONAL RECOMMENDATION SYSTEM: EXPERIMENTATION WITH A VOTING-BASED EVALUATION MODE

Our research work falls within the context of the use of recommendation systems for the personalization of contents in e-learning environments. We present in this paper the results of a second experiment that was recently conducted to evaluate a hybrid recommendation approach in an online learning environment. The approach consists of combining the two approaches of content-based filtering and collaborative filtering to improve the relevance of the educational resources recommended to learners. A first experiment was carried out in 2019 and gave convincing results, which led us to repeat a second experimentation in order to confirm the results on the one hand, and on the other hand, to modify the way learners evaluate the resources by transforming the "like" by a voting mode from one to five, in order to verify whether this will bring an improvement in the recommendations. This second experiment was also an opportunity to integrate an engine that guides learners' searches by adding criteria relating to their preferences and to check their satisfaction with the use of this engine. The results were globally positive.