Opposition-Based Genetic Algorithm for Community Detection in Social Networks

This paper proposes an improvised algorithm named modified crossover opposition-based genetic algorithm (MCOBGA) for community detection with the help of genetic algorithm (GA) to discover community structure in social networks. The paper deploys modified crossover and opposition-based initialization along with GA to improve the quality of the community structures. Initialization of the population through opposition-based learning ensures the improved selection of initial population, whereas modified crossover transmits information for improved community structure. The evaluations of proposed algorithm have been done on real-world networks. The experimental results show that MCOBGA has very competitive performance compared with GA with vertex similarity applied to community detection which has been the most similar approach to the proposed algorithm. Experimental results not only demonstrate improvement on convergence rate of the algorithm, but also communities discovered by proposed algorithm (MCOBGA) is highly inclined towards quality, compared to its counterpart. In this paper, we have focused on the community detection problem in the domain of the social network. Community detection is a very basic and hot research problem in complex networks. We have employed the genetic algorithm with modified crossover, opposition-based learning, and matrix encoding technique. We can use this technique in agriculture, the health sector, and market data analysis also.