Research On Community Clustering In Complex Networks Based On MapReduce Clustering Algorithm

Such as transportation networks,social networks and other complex systems in the real world can be modeled as a complex network composed of a plurality of constitute the "community" or "clusters",reveals its community structure in order to understand the network structure and network behavior analysis by researchers attention.Community clustering is to find the community structure in complex networks,and then extract the important information contained in the community structure.However,with the growing of mobile Internet,social network,internet of things and other complex network in recent years,the network scale grows unceasingly.The traditional stand-alone mode of community clustering method has been unable to meet the needs of large-scale complex network analysis,how to deal with the large-scale complex network community cluster becomes a hot spot in current research.The thesis of the research work of complex network clustering based on MapReduce clustering algorithm to the problem of complex network clustering has been proposed as follows:Firstly,a complex network clustering method based on neighborhood search clustering algorithm is proposed to solve the cluster problem of complex network community.The algorithm selects the clustering center through the neighborhood search control strategy and overcomes the randomness and limitation of the traditional clustering algorithm during the time of selecting the clustering center,so as to realize the better community clustering effect.The experimental results show that the proposed algorithm has a high detection accuracy in the relatively small complex network.Secondly,with the scale of complex networks increasing,the traditional stand-alone mode has not been able to satisfy the problem of large-scale network clustering,so this thesis proposes a complex network clustering method based on neighborhood search clustering algorithm with MapReduce.In order to realize the cluster processing of large-scale network community after MapReduce,this method includes data preprocessing,calculating the shortest path of nodes,calculating neighborhood density,selecting clustering center nodes and clustering.Finally,the duster of Hadoop is designed as an experimental platform,and the experimental results show that with the increasing of the network scales,the algorithm based on MapReduce has obvious advantages in the execution speed compared with single machine which shows high accuracy and reliability.