A framework to mine communities using nature inspired algorithms

Natural intelligence heuristic techniques have demonstrated their capability to provide acceptable solutions to many real life complex problems. Their potential to mine communities from complex networks has been successfully tested by many researchers. With the growing rate of development of new robust and efficient nature based algorithms, a strong need is felt for a generalized framework to evolve communities which can accommodate any existing or new nature based algorithm. This paper proposes a framework for applying natural intelligence based optimization strategies to detect communities using any nature inspired algorithm. The proposed framework can serve as an abstraction for evaluating a set of new/existing nature based methodologies for detecting communities at varied level of optimalities in order to select the most efficient algorithm. Extraction of communities at varied optimality levels by considered algorithms can reveal multiple grouping patterns prevailing in the complex network and can help in planning strategic decision. The framework consists of four prominent, independent phases: The Analysis, Initialization, Evolution and Result Generation Phase. Evaluation of any new/existing metaheuristic or evolutionary algorithm for community detection may be simply done by modifying the Evolution Phase. The framework thus provides an easy to use flexible platform for use by researchers in the domain of community detection.