Using maximal clique identification techniques to design an effective graph-based social recommender system

Social recommender systems help users find their required information much faster. The main challenge in these systems is accuracy and quality of the recommendations they provide. Application of more appropriate algorithms anddatastructurescanleadtoimprovementsinbothoftheabove-mentioned criteria. In this paper, first we reconsider the ideas behind finding maximal cliques in the graph theory, and then we argue on inspiring from it to design a new social recommender system. We represent a social network by a graph in which vertices are the users and edges are their connections and communications. Thisresearchworkshowshowidentifyingallmaximalcliquesinsuch a graph lets us model the system to find user‘s similarity interests and needs. Formal analytical investigations show that the proposed idea leads to significantimprovementsinaccuracyandqualityoftheprovidedrecommendations.