Attribute network embedding and parameter-free clustering-based Bilibili user community discovery method

Abstract

The invention relates to an attribute network embedding and parameter-free clustering-based Bilibili user community discovery method, which comprises the following steps of: constructing an attributenetwork G according to the existing Bilibili user data, converting an adjacent matrix and an attribute matrix into an embedding vector H by utilizing an attribute network representation framework, operating a k-means algorithm on the H, calculating the minimum intra-group variance within T times under different k values, calculating the maximum curvature index and the second maximum curvature index with the highest occurrence frequency in I times to obtain corresponding k values ka and kb, respectively operating k-means algorithms with IT secondary community number ka and community number kb,calculating the maximum modularity Qa and Qb of the clustering result, obtaining the k value corresponding to the larger value in Qa and Qb as the community division number, and finally realizing community discovery by using the k-means clustering algorithm. According to the invention, the parameter-free algorithm is used to replace manual assignment of the number of communities, human factor interference is reduced, and the accuracy and stability of community discovery are improved.

Description

technical field [0001] The invention relates to the field of network science, in particular to a Bilibili user community discovery method based on attribute network embedding and non-parametric clustering. Background technique [0002] With the rapid development of science and technology, the Internet has greatly promoted the development of all aspects of society and changed all aspects of people's lives. The video social network represented by Bilibili has become an important part of people's daily entertainment and leisure. Bilibili is a popular website among young people. It has both the functions of a video website and a social networking site. Users can post, watch videos, and comment on videos on the website, and at the same time follow their favorite bloggers , you can like, coin, and bookmark your favorite videos. Therefore, it has attracted a large number of young users. The user community on Bilibili is a non-physical network of social networks. In this network, ...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the defects of low accuracy and high complexity in the current large-scale attribute network community discovery method, this invention proposes an efficient Bilibili user community based on attribute network representation learning and non-parametric clustering The discovery method first uses the attribute network representation learning framework to calculate the embedding vector of each user, and then uses the non-parametric clustering algorithm based on curvature and modularity to determine the number of community divisions and realize community discovery, improving the accuracy and accuracy of the community discovery algorithm. efficiency

Images