A Community Detection Algorithm Based on NSGA-II
Abstract: The community
detection problem is modeled as multi-objective optimization problem, and a classic
NSGA-II (nondominated sorting genetic algorithm) is adopted to optimize this
problem, which overcomes the resolution problem in the process of modularity
density optimization and the parameter adjustment in the process of general
modularity density optimization. In this case, a set of Pareto solutions with
different partitioning results can be obtained in one time, which can be chosen
by the decision maker. Besides that, the crossover and mutation operators take
the neighborhood information of the vertices of networks into consideration,
which matches up with the property of real world complex networks. The graph
based on coding scheme confirms the self-adjustment of the community numbers,
rather than sets up in advance. All the experiment results indicate that
NSGA-II based algorithm can detect the construction of community effectively.
Keywords: complex Networks,
community detection, modularity density function, NSGA-II, multiobjective evolutionary
algorithm
Author: Lishuo Zhang, Qi Wang
Journal Code: jptkomputergg160058
