Research Of Community-seeking Algorithms Based On Service Network

Seeking communities is aimed at revealing the real cluster structure of complex network. The service network is a network set which has formed as the result of interaction between lots of open service. With services combining spontaneously to meet users’ requirement driven by business needs many service clusters will emerge. Seeking and make use of these clusters is helpful to analysing service system on the whole and to looking for service groups with strong behavior correlation.It has important significance and application value for building service network.The thesis carry on research about community-seeking algorithm of service network based the theory of community structure mining.It includes two parts as follows:first, it provides the new link similarity definition and measure method between nodes of service network. Community-seeking algorithm based link similarity is given further. The thesis proves that the algorithm is reasonable when it is used on the service network by experiments on some data sets and uses it in service structure network to verify the effectiveness of the algorithm compared with GN algorithm.Secondly after summarizing of the dynamic executive behavior of service network the thesis give the community-seeking algorithm on weighted network which is verified on the classic network. The algorithm is applied to real service executive network and proved correct by analysing the obtained service communities.The experiments show that the algorithms given in this thesis based on service network is reasonable and effective.