Research On Detecting Community Evolution Of Network

With the advent of the era of big data, how people from such huge data miningvaluable information is particularly important. In many cases these massive data canbe abstracted into complex network problems. During the study of complex networks,we found that the community analysis and research-related issues has always been ahot issue. Scholars from the simplified static community discovery research began,gradually developed into the study of dynamic complex networks. Due to thecomplexity of network topology changes dynamically over time, so the study ofdynamic network, not only need to be a dynamic community to explore, but also beable to monitor the course of its evolution, which contribute to the improvement ofpeople characteristics and understanding of the law of development of the entirenetwork structure has important theoretical and practical significance. Therefore, thisarticle will find a dynamic online community to monitor its evolution as the twostudy sites.We need to find the community befor we monitor the evolution.(1) In a dynamic online community discovery, the paper found that thinkingthrough analysis based on incremental dynamic community found time not only hashigh efficiency, but also has a good consistency around the community. Therefore, weuse these algorithm ideas and learn from cluster analysis algorithm ROCK thinking,the use of the number of common neighbor similarity that measure to change part ofthe community re-division of the increment. To overcome the traditionalcommunity-based incremental discovery algorithm requires a fixed number ofcommunities can not find the problem the new community.(2) Monitoring the evolution of the community, the paper found through analysisbased on the idea of community to monitor the evolution of the central node in thesame time have a higher efficiency have good accuracy. Therefore, we use thisalgorithm thinking, and using a mechanism similar to the vote at the central node selection process, the original algorithm to solve the problem you need to set thethreshold to improve the usability of the algorithm.This paper uses the two data sets for each algorithm is proposed experiment.Analysis of the experimental results show that the proposed algorithm based onincremental time complexity of the original algorithm similar situation can be foundin the effect of better communities, but also to discover new communities. In addition,the paper is really the evolution of community monitoring data sets verify theeffectiveness of the monitoring and analysis of algorithms evolution of thecommunity.