Research on of overlapping community detection algorithm based on tag influence

Because of the overlapping community detection algorithm is random and easily forms the monster community, an overlapping community detection algorithm OCDA_TI based on tag influence is proposed in this paper. Firstly, the concept of subordinate degree that represents the ascription degree for this community vertices in different communities is defined in the algorithm; secondly, for the problem that the attraction between tag vertex will weaken according to the tag propagation distance increases, tag score and the attenuation factor is described. In order to avoid the problem of random selection of same label influence, the similarity measure is defined; thirdly, the calculation method of tag influence value and the termination condition of tag transmission are given based on the subordinate degree function and attenuation factor. Considering that the network structure is difficult to determine the attenuation factor, the propagation distance parameter is introduced, which combines the modularity increment maximum; Finally, the testing of the OCDA_TI algorithm in different data sets, the experimental results show that the proposed algorithm has good stability, and the quality of community detection is superior to the typical overlapping community detection algorithms.