Font Size: a A A

Research On Complex Network Community Detection Algorithm Based On Affinity Propagation

Posted on:2018-06-25Degree:MasterType:Thesis
Country:ChinaCandidate:G B SunFull Text:PDF
GTID:2310330539975495Subject:Computer system architecture
Abstract/Summary:PDF Full Text Request
Researches have shown that the community structures are generally existing in the complex network.The nodes within the community are closely connecting with each other,but the nodes between the communities are sparsely connecting with each other.The community structures within complex networks have very important theoretical and practical value,which can be used to help us understand the functions of complex networks,discover the hidden rules in complex networks and predict the behavior of complex networks.The topic mainly focuses on the researches of the community detection on complex networks and the affinity propagation algorithm,which includes the three following aspects:Firstly,we proposed a community detection algorithm based on structural similarity affinity propagation(SS-FAP).The first,the algorithm selected structural similarity as a similarity measure between nodes,and applied an optimized method to calculate the similarity matrix.The second,the algorithm made the similarity matrix as an input,and used a fast affinity propagation algorithm to cluster.The last,the algorithm got the final community structures.The experimental results showed that the SS-FAP algorithm had better community detection ability and can detect higher quality community structures on the LFR simulated networks or real networks,Secondly,we proposed a community detection algorithm based on modularity affinity propagation(MAP).The main idea of MAP was introducing the modularity function into the iterative process of AP algorithm to optimize the community detection result.The experimental results showed that the MAP algorithm,compared with traditional LPA algorithm,FN algorithm,BGLL algorithm and AP algorithm,can more effectively detect the community structure in the network.Finally,we implemented a community detection algorithm prototype system.The prototype system implemented SS-FAP algorithm,MAP algorithm,LPA algorithm and four evaluation criteria,which are NMI,FM,Accuracy and Modularity,and used force-guided layout algorithm to display networks.
Keywords/Search Tags:complex network, community detection, affinity propagation, structural similarity, modularity
PDF Full Text Request
Related items