A Recommendation Approach Based on Community Detection and Event Correlation Within Social Learning Network

Abstract

The social learning network is regarded as one of the most widespread types of online learning due to its collaborative and interactive properties. Recommendation systems contribute significantly in delivering relevant contents on social networks. However, learners are not consistently engaged. It is inescapable to develop a strategy for providing recommendations to meet system requirements for non-considerably active periods and to interrelate all events. To mitigate this problem, we propose an approach to fill the information gap and make recommendations more reliable. Our approach therefore introduces community detection and the correlation between data related to all events carried out by learners: consultations, information sharing, discussions, researches, etc. It is based on (a) detecting communities of learners who interact more intensely with each other and share common interests, and (b) calculating recommendations based on communities detected and computed correlations between data associated to learners’ events. Our perspective allows us to jointly embed two crucial guidelines, which are the event correlation and community detection.