Fuzzy Cognitive Map-Based Genetic Algorithm for Community Detection

摘要

One of the most elemental operations concerning the analysis of properties of a network is community detection. It is the process of decomposition of a given network into groups of densely connected nodes that tend to share some similar properties. A wide variety of algorithms to identify the communities in complex networks exists. In this paper, an intelligent genetic algorithm (GA)-based approach to identify communities has been proposed. The efficiency of the solution that resulted from the genetic algorithm depends on the setting appropriate values for the various parameters involved. As a means to reduce the convergence time of the genetic algorithm, a fuzzy cognitive map (FCM) is used. The knowledge derived from the FCM is used to populate the initial population reducing the randomness of the algorithm. The potency of the algorithm is evaluated on various weighted and unweighted benchmark networks.