Social Network Clustering

As we know, the datasets related to social networks are increasing. There are different procedures to analyze these types of datasets; one of these procedures is clustering which makes communities of social data. Random walk is a process which can find communities in a network, in other words when a random walk is used, it scans the nodes in some steps; it begins with an initial node and based on a random process progresses to neighboring nodes. In this paper an algorithm is proposed which aims to finding communities in a way that modularity factor increases, for this goal, random walks with random local search agent are combined. Experimental results show that the proposed method gives better modularity in comparison with other algorithms.