Overlapping community discovery method based on node influence

Abstract

The invention provides an overlapping community discovery method based on node influence. On the basis of a k-shell decomposition algorithm, three classes of factors, including the global influence and the partial influence of a node and the contribution degree of a neighbour node are comprehensively considered on a full scale. On the basis of the influence, nodes are updated in a descending order, the propagation characteristics of social network information are conformed, and algorithm randomness is lowered. When an affiliation coefficient is calculated, the impact of the influence of the neighbour node is considered, and the own affiliation tendency of the node and the own influence of the node are considered. On an aspect of setting an overlapping threshold value and the like, the node influence is fused, a reasonable threshold value is set for each node. The randomness of an original algorithm can be lowered, and a high-quality community structure is mined.

Description

Technical field [0001] The invention relates to an overlapping community discovery method. Background technique [0002] In nature and human society, there are many real networks that can be described as complex networks. For example, the information flow network, the metabolic system network and the protein network in the biological world. The most common complex network is a social network formed by individual users and their relationships (such relationships may be relatives, colleagues, friends, etc.). In short, a social network is the network between users, users and users. A network of relationships. People hope that through effective and quantitative data mining and analysis, the hidden information and laws can be revealed, which greatly promotes the research boom of social networks. [0003] With the deepening of research, people gradually discovered that social networks not only have small-world and scale-free characteristics, but also have network cluster structure cha...

AI Technical Summary

Problems solved by technology

[0020] The purpose of the present invention is to provide an overlapping community discovery method based on node influence, which can solve the problems of the randomness of the COPRA algorithm in the label propagation class overlapping community discovery algorithm and the lack of consideration of node influence in the calculation process.

Images