Community detection method based on similarity and dissimilarity constraint semi-supervised non-negative matrix factorization

Abstract

The invention discloses a community detection method based on similarity and dissimilarity constraint semi-supervised non-negative matrix factorization. The method mainly comprises the following steps: 1) converting edge connection data of a community complex network into an adjacent matrix A; 2) randomly extracting link information according to a ratio to obtain a priori subset, and then respectively constructing a Must-link matrix M and a Cannot-link matrix C; 3) setting a maximum number of iterations maxiter and a non-negative feature dimension k, and randomly initializing a basis matrix W and a coefficient matrix G; and 4) obtaining a category label of the community according to the updated coefficient matrix G. The invention provides a semi-supervised community detection algorithm based on non-negative matrix factorization, introduces prior information of mus-linked and cannot-link, respectively constructs similarity constraints and dissimilarity constraints between data samples, improves the characterization capability of community network nodes, and improves the detection accuracy through experimental verification and analysis of a real community network data set. Through experimental verification and analysis of a real community network data set, the precision of community detection is effectively improved.

Description

technical field [0001] The invention belongs to the technical field of complex network data processing, mainly studies the community detection problem of undirected unweighted network, and relates to a community detection method based on similarity and difference constraints semi-supervised non-negative matrix decomposition. Background technique [0002] In real life, complex network structures are ubiquitous. In many fields, data usually exists in the form of network structures, such as social networks, citation networks, biological networks and technical networks. A network is a structure with modules or communities. In the network, each entity is regarded as a node, and the connection between nodes is regarded as an edge. In this way, the network can be modeled as a graph. Each module or community is similar to a subgraph, and the connection between the internal nodes of the community will be closer, and the connection with the external nodes of the community will be rel...

AI Technical Summary

Problems solved by technology

[0004] In the process of community detection, due to the sparsity and noise of the community structure, only relying on topological information is not enough to accurately find the community structure, and usually some potentially useful prior information can be obtained, so how to combine the topology and prior information of the network information to improve the accuracy of community detection is very challenging

Images