Influence community search method and system based on community retrieval

The invention discloses an influence community search method based on community retrieval. The method comprises the following steps: setting an initial threshold value of a weight through a heuristicmethod to screen a network to generate sub-graphs, thereby reducing the scale of the sub-graphs and achieving the purpose of reducing the calculation cost; judging the connectivity of the key nodes through the spanning tree on the basis of the subgraphs so as to count the number of communities; carrying out incremental iteration on the threshold values which do not meet the query conditions untilthe threshold values meet the query conditions; and finally, outputting the community with the maximum influence progressively according to the query condition, so that the repeated enumeration of thenumber of the communities is avoided. According to the method, the defects of the existing community retrieval technology are overcome, but the advantages of the existing community retrieval technology are reserved, so that the community retrieval technology becomes perfect: firstly, repeated calculation is effectively reduced through a progressive output result, and the calculation cost is saved; secondly, the influence of the nodes with the same weight on community retrieval is fully considered, the limitation condition of the weight is broken through while the retrieval performance of thenodes is not influenced, and the application range of the nodes is expanded.