Community Detection in Social Media Network with Maximum Modularity Using Girvan-Newman Algorithm

Social networks are formed from interactions of peoples. Measuring the degree of those relationships requires interpreting connectivity of vertices and extracting information from it. Generally individuals form smaller sub-communities in those networks. Identifying those communities by determining sizes of cliques is a challenge and there are numerous solutions in the literature for this problem. In this study we reviewed Girvan-Newman community detection algorithm and applied it on a real life social network obtained from Twitter data. Friendship relations of students of four different universities were used to form the network. A connected graph is generated from this data set in which the students are represented as vertices and followership relations of the students formed the edges of the graph. Since those universities are geographically close to each other, the graph consisted of different link connections among those four clusters. Then community clusters were detected in this connected graph by using Girvan-Newman community detection algorithm