Font Size: a A A

The Methods Of Overlapping Community Discovery And Node Evaluation In Social Network

Posted on:2018-06-26Degree:MasterType:Thesis
Country:ChinaCandidate:H F LingFull Text:PDF
GTID:2310330569486465Subject:Computer technology
Abstract/Summary:PDF Full Text Request
With the popular of social softwares such as QQ,Facebook,microblogging,WeChat and so on,social software has become an essential part of people's life.Based on the analysis of the existing problems in community discovery algorithms and node evaluation algorithms,this thesis mainly focuses on the study of overlapping communities discovery in social networks.The main works are as follows:1.In view of the problems that high parameter dependency,excessive degree of overlap,uncertain of the stability of multiple runs in overlapping community discovery algorithms,this thesis presents a method for calculating the degree of node attribution,and the algorithm of overlapping community discovery algorithm is designed according to the idea of local optimal expansion of cohesion clustering.First,the initial core community is constructed with the node with the highest weight and its neighbor nodes.Then the core community is extended by node attribution until the termination condition of the algorithm is satisfied.Finally,by the comparison with the existing algorithms,the results show that the improved algorithm has much better performance compared with the other algorithms,and the community structure is more reasonable.2.For the unreasonable average division of weights of traditional PageRank algorithm in the social network,this thesis proposes a method to calculate the authoritative degree of nodes,and improves the shortage of PageRank algorithm,the improved algorithm divides the weights according to the weights of the nodes.By comparing with the existing classical node evaluation algorithms,the experimental results show that the association between nodes of the improved algorithm is stronger and the evaluation results are more reasonable than the original ones.3.On the basis of the above researches,this thesis designs and implements a prototype system of overlapping community discovery and node evaluation in social network,and introduces the framework,function modules and development process of the prototype system in detail.The system has realized the function of overlapping communities discovery,nodes evaluation and analysis of communities structure and so on,and has some practical value.
Keywords/Search Tags:social network, community discovery, overlapping community, node importance, node evaluation
PDF Full Text Request
Related items