Enhancing Community Detection Using Maximal and Maximum Cliques into Hierarchical Algorithms

Nowadays, social networks have expanded rapidly. People tend to build clusters based on relationships with others. One of the most crucial fields that aids in understanding and analyzing the structure of huge and complex networks, is communities' detection. The goal of community detection is to reduction the application-generated graph into smaller communities with comparable nodes. The counting of cliques in a larger network is a fundamental problem in graph theory. The current study concentrates on a structural network which only depends on links between users to detect communities. the proposed model was developed to enhance the Leiden community detection method. We combined the Leiden algorithm with the Maximal and Maximum cliques, where the cliques were first calculated, then the Maximal and the Maximum clique were extracted from these cliques, then the maximal and maximum clique were entered into the Leiden algorithm. Finally, the proposed model can enhance the quality function of Modularity for Leiden community detection method. Several experiments have been done on Facebook and Twitter. The results have shown improvement in the Leiden quality for Modularity,Where the modularity have increased in Facebook Dataset from 0.79 to 0.81 and also in Twitter it increased from 0.76 to 0.79.