Affect of Varying the Similarity Threshold in Leader Similarity Based Community Detection

In the present scenario community detection is significant for extracting the community structure of a given network. It is very important in understanding and evaluating the formation of large and complex networks. It provides us the competency to look into group-level. There are numerous algorithms that have been put forward to study the community structure in a network and had been applied to various domains. They help in classifying the nodes on the basis of their functions in account of their positions in the communities. In this paper our focus is to study the sensitivity of the similarity threshold extending the leader similarity based community detection algorithm i.e. first of all leader is selected from the network using Eigen vector centrality than on the basis of the similarity threshold the followers are selected and we get the community in the network. We also analyzed how the variation of the similarity threshold helps us to evaluate the size of the community.