Cross-social network user identification method based on full-view features

Abstract

The invention discloses a cross-social network user identification method based on full-view features. First, multiple social networks are divided into communities and reference points are initialized. Then, iteratively carry out the following three steps: (1) Use the reference point to calculate the full-view feature of the unrecognized user to calculate the similarity between users; (2) Use the improved stable marriage matching algorithm to complete the user identification work; ( 3) For newly identified user pairs, update the set of reference points according to features such as community centrality. The above three steps are repeated continuously until the reference point set is no longer updated, and a matching anchor link user set is obtained. Adopting the cross-social network user identification method of the present invention, on the one hand, the global position of the user on the social network is considered, and the accuracy rate and recall rate of the user identification method are improved; in addition, through the iteratively adjusted identification strategy, multiple The problem of correct identification of user pairs with similar similarities also avoids the problem of cold start.

Description

technical field [0001] The invention belongs to the field of data mining and data integration of social networks, and mainly relates to a cross-social network user identification method based on full-view features. Background technique [0002] With the development of the Internet, more and more people have established various virtual accounts on the Internet. Different from traditional SMS (Short Message Service) and other applications, social networks, as a product of the WEB2.0 era, focus on social attributes and provide people with a wealth of social services, such as using social networks to share news, transfer knowledge, publish topics, etc. . Most of them communicate with friends through multiple social networks, but different accounts of users are not related to each other because they are distributed on different social networks. If these online relationships can be merged into a single environment, it can help users keep in touch, and also provide a way for user...

AI Technical Summary

Problems solved by technology

[0005] The existing hybrid social network user matching research mainly considers the processing of information characteristics of social networks, and there are some shortcomings: first, it is the cold start problem

If the heuristic algorithm is used blindly to solve the problem, the calculation of the initial point will be separated from the structure, which will have a certain impact on the subsequent calculation.

Second, only local structural information is considered

Existing methods only consider local information such as the neighbors of the two or the places associated with the two.

Images