基于动量粒子群优化的社会网络分析

Analysis of Social Networks Based on the Momentum Particle Swarm Optimization

针对社会网络分析中的社区发现问题,在原有的粒子群优化算法的基础上,提出了一种基于动量粒子群优化算法,并且将此算法应用于社会网络分析中的社区发现研究中,提出了一种自适应社区发现方法.利用Newman提出的模块度作为适应度函数,在优化过程中自动获取社区数目,在Karate网络上的实验结果表明,所提出的算法能够有效地进行社区预测,并且获得了较高的预测精度.

In terms of social network analysis, a new momentum particle swarm optimization algorithm based on the original thoughts of PSO was proposed. By this algorithm, the social network analysis was applied to solve community detection problems. An adaptive community discovery algorithm based on momentum particle swarm optimization was further proposed. By using Newman＇s modularity as fitness function, the number of communities in the optimization process was obtained. Experiments on Karate network showed that the algorithm could effectively predict the community and obtain perfect prediction accuracy.