Community detection in complex networks using structural similarity

These days, community detection is an important field to understand the topology and functions in the complex networks. In this article, we propose a novel Community Detection Algorithm based on Structural Similarity (CDASS) that executed in two consecutive phases. In the first phase, we randomly remove some low similarity edges. Therefore, the network graph is converted into several disconnected components that are considered as primary communities. In the following, the primary communities are merged in order to identify the final community structure close to real communities. In the second phase, we use an our identified evaluation function to select the best communities between overall random generated partitions. Finally, we evaluate CDASS algorithm using several scenarios extracted from artificial and real networks. The results, obtained from simulation with these scenarios, show that proposed algorithm detects communities with high accuracy close to optimal case and is applicable in the large and small network topologies.