Associated user identity recognition method based on social network topological graph

Abstract

According to most current methods, a social network is embedded into a low-dimensional vector space, and then users are aligned into the low-dimensional space. However, since the social network is extremely complex and huge, the social network is easily influenced by error propagation and noise of different neighbors in the network embedding process. On the basis, the invention provides an associated user identity recognition method based on a social network topological graph, which comprises the following steps: firstly, forming an ego network of a user (namely, extracting a local network formed by a section of neighbors of the user), then extracting a user node sequence by using random walk, and then learning low-dimensional vector representation of the user by using a natural language model framework; and finally, the training matrix maps the two social networks to the same feature space for alignment. According to the invention, interference caused by high-order neighbors can be avoided by using the ego network, so that the node embedding result can be improved, and the association accuracy can be improved.

Description

technical field [0001] The invention relates to the technical field of multi-interaction network data analysis and mining, in particular to an associated user identity identification method based on a social network topology map. Background technique [0002] Associative user identification, which aims to discover the correspondence between the different identities of the same user in multiple social network platforms, is a key technology in the field of data analysis and mining of multiple social networks, and has a wide range of commercial application requirements. There are important applications in security and personal recommendation. [0003] Most of the current methods are based on DeepWalk (Perozzi B., AI-Rfou R., Skiena S. DeepWalk: Online learning of social representations [C] / / Proceedings of the 20th ACM SIGKDD International Conference on Knowledge discovery and datamining. New York: ACM Press, 2014:701-710.), which draws on Word2vec (Mikolov T., Sutskever I., Ch...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to propose a method for identifying associated user identity based on a social network topology graph, which is used to solve the problem that the existing method introduces too many high-order neighbors (that is, those that are not directly connected to the node) when embedding neighbor nodes. Noise from other nodes), which leads to the technical problem of low recognition accuracy

Images