Multi-granularity community discovery method based on improvement density peak value

Abstract

The present invention claims a multi-granularity community discovery method based on the improvement density peak value. With the method, firstly, the disadvantages of the density peak clustering algorithm are tackled that the cluster center is difficult to determine and the classification is error-prone, and the cluster center can be accurately found; secondly, according to the intermediate results of density peak value clustering, a global community topology structure map is constructed, according to the defined granularity decomposition rules, the granularity layer is automatically dividedfrom coarse to fine on an initial global community map, and the initial global community topology structure is divided by a decomposition mechanism to obtain multiple independent community structureson the fine granularity level, so that the problem solving space is enabled to be simple; and finally, the optimal division of the social network community structure in the optimal granularity level space can be formed. The method reveals the internal hierarchical relationships of social networks to a certain extent, and can quickly and accurately discover the stable hierarchical structure of thenetwork.

Description

technical field [0001] The invention belongs to the fields of social network analysis, granular computing, cluster analysis, etc., and in particular relates to a community structure discovery method based on improved density peak algorithm and multi-granularity thinking. Background technique [0002] The community structure is a reflection of the modularity and heterogeneity of the network, which means that the real network can be regarded as a combination of many different types of nodes, such as the circle of friends in the interpersonal network, the related papers on the same topic in the citation network, the metabolism or Functional subgroups in protein networks and more. In-depth study of the community structure of the network not only helps to reveal how the intricate real network is formed by many relatively independent but interrelated communities, but also enables people to better understand the structure and functional characteristics of different levels of the sy...

AI Technical Summary

Problems solved by technology

At the same time, there are some problems in some current community discovery algorithms. It is difficult to deal with large-scale complex networks, the community structure is unstable, and there are large errors in the method of specifying the number of communities in advance, etc.

Images