Community structure identification method and device based on network embedding

Abstract

The invention discloses a community structure identification method and device based on network embedding, and relates to the field of artificial intelligence and complex networks. The method is usedfor solving the problem that low-dimensional representation of an existing node cannot capture community structure information and structural similarity. The method comprises the following steps: determining a first-order similarity matrix and a second-order similarity matrix of a network according to an adjacent matrix of the network; determining a Dice's similarity matrix of the network according to the two nodes which are neighbor nodes of each other; determining a structural similarity matrix of the network according to the derived sub-graph of the network and the number of the self-orbits; determining a final similarity matrix of the network according to the first-order similarity matrix, the second-order similarity matrix, the Dice's similarity matrix and the structural similarity matrix; obtaining a low-dimensional representation matrix of the node according to the non-negative matrix decomposition, the final similarity matrix, the community member guidance matrix and the community low-dimensional representation matrix; and clustering the low-dimensional representation matrixes of the nodes according to the k-means to obtain community division of the network.

Description

technical field [0001] The present invention relates to the fields of artificial intelligence and complex networks, and more specifically to a method and device for identifying community structures based on network embedding. Background technique [0002] Some complex relationships in the real world can be described by networks, entities in the network can be represented by nodes in the abstract network, and links between entities can be described by edges. Using complex networks to model the real world is a very effective method. Complex networks use scientific research methods to present data in the real world in an easy-to-understand and apply way, and because of this, the current research on complex networks has received extensive attention. In a complex network, the community structure in the network is one of the most important features, and it is an important method to reveal the structure, function and dynamic changes of the entire network. In addition, identifying...

AI Technical Summary

Problems solved by technology

[0008] Embodiments of the present invention provide a community structure identification method and device based on network embedding to solve the problem that the existing community structure identification method does not consider the relationship between nodes in the network, resulting in that the low-dimensional representation of nodes cannot capture the community structure Questions of information and structural similarity

Images