Ranking Of Nodes And Community Detection In Mobile Social Network

Complex network is an another expression way of system in real world and has some basic statistical characteristics.Research of complex network gives people a better understanding and grasping about system in real world and has great theoretical and practical values.Ranking of nodes may be used to analyze importance of every node in social network.Core nodes could be found through these analysis and other nodes may be influenced or changed through core nodes.Community detection could be used to analyze topology,composition and hide laws of complex network.This thesis is based on theories and methods in data mining and does some deep studyings of ranking of nodes and community detection in mobile social network.Main tasks are as follows:1)Discussion about purpose and meaning and introduction to main methods and research status of ranking of nodes and community detection are included in this thesis.2)Based on PageRank algorithm,PhonePageRank algorithm is proposed which could be used to quantize importance of nodes in mobile social network.Effectiveness of PhonePageRank algorithm is proofed using it to do some influencing anslysis of half-year communication datas which are gotten using questionnaire by a research group.3)PPRLPA algorithm is proposed based on PhonePageRank and LPA(label propagation algorithm).Some influential seeding nodes which are dependent on values of nodes are endowed labels.Label propagation is achieved from seeding nodes.Effectiveness of PPRLPA algorithm is proofed based on communication networks of a research group and some other real data sets.4)Based on real demands,demanding analysis,system designing and system developing of thread chains analyzing system in mobile social network is developed and realized using data mining of mobile social network as core algorithm.Software foreground is achieved using Windows Form packages.Database is implemented using SqlServer2008.Software background codes are realized using MVP frame separating interface,functional logics and data models.There are four functions which could be used by users:data establishing,data analyzing,basic datas and system managing.Results of simulation indicate that system developed has great data mining efficiency and high performance and user demands could be nicely met.