Community detection using local neighborhood in complex networks

• Local neighborhood ratio is proposed in defining community structure.
• Introduce the average neighborhood ratio as threshold in associating nodes to a particular community structure.
• Presents related theorem and its proofs about the neighborhood ratio in defining community structures in complex networks.
• A new method is presented in detecting community structures in complex networks.

It is common to characterize community structure in complex networks using local neighborhood. Existing related methods fail to estimate the accurate number of nodes present in each community in the network. In this paper a community detection algorithm using local community neighborhood ratio function is proposed. The proposed algorithm predicts vertex association to a specific community using visited node overlapped neighbors. In the beginning, the algorithm detects local communities; then through iterations and local neighborhood ratio function, final communities are detected by merging close related local communities. Analysis of simulation results on real and artificial networks shows the proposed algorithm detects well defined communities in both networks by wide margin.