Social network community discovery method and device based on improved label propagation, and medium

Abstract

The invention discloses a social network community discovery method and device based on improved label propagation, and a medium, and the method comprises the steps: constructing an undirected networkthrough employing users in a social network as nodes and the attention relation between the users as a connection edge; calculating the betweenness of each node in the network, and then sequentiallystoring all the nodes in the vector according to the betweenness to serve as a new node set; calculating a Salton similarity index between the nodes; allocating a different label to each node in the network, and iteratively updating each node label based on the similarity matrix until each node label does not change any more; and dividing the nodes with the same label into the same community. According to the method, the network structure is mined more fully, the labels are distinguished more finely, and the convergence rate of the algorithm can be increased.

Description

technical field [0001] The invention relates to the field of social network and network science, and specifically designs a method, device and medium for discovering social network communities based on improved tag propagation for typical microblog networks in social networks. Background technique [0002] In recent years, social networks have played an increasingly important role in people's lives. From foreign Facebook and Twitter to domestic microblogs and blogs, they have greatly affected people's daily life and communication methods. Among them, the number of users of Sina Weibo has achieved a blowout growth within a few years, with an average daily active user of more than 50 million. Sina Weibo, as an open speech platform and a convenient communication channel, realizes the rapid dissemination of topics and news by building a user attention relationship network, which greatly improves the public's participation in hot topics. In the Weibo network, users can become p...

AI Technical Summary

Problems solved by technology

However, most of the existing label propagation algorithms do not consider the full use of the network structure. Random selection methods are often used in the process of label updating and selection, resulting in unstable community structures, and multiple different division results may be obtained in the same network.

Images