SIG-CLA: A Significant Community Detection based on Cellular Learning Automata

Abstract

Detecting community, as the fundamental task for the study of the network, reveals a hopeful approach to investigating the functional and topological properties of real networks. Recently, several algorithms for detecting community have been introduced from various perspectives. A typical algorithm which is noticed by many researchers in this scope is Modularity optimization. These algorithms significantly restricted in resolution limits that they may miss detecting communities that are less than a particular size. The paper presents an algorithm with the aid of ICLA (irregular cellular learning automata) for the detection of community structures (called SIG-CLA) in social networks. In the SIG-CLA, the ICLA is formed by the input network, and the communities are detected by communicating with both the local environment and the global environment via the significant function in the ICLA. Promising experimental results are presented to confirm the advantages of SIG-CLA with respect to Modularity and NMI.