A method and system for cross-domain identity association based on multi-relational representation

Abstract

The present invention relates to a method and system for cross-domain identity association based on multi-relational representation, respectively acquiring entity feature vectors of two entities to be associated with identity, wherein the two entities to be associated with identity belong to two different domains respectively; The respective entity feature vectors of the entities to be associated with the identity and the final cross-domain identity association model are obtained to obtain the association similarity between the two entities to be associated to represent the cross-domain identity association, that is, to determine the association between the two to be identified. Whether the entity is the same entity, the speed is fast, the precision is high, and it has more practical value in the industry.

Description

technical field [0001] The invention relates to the technical field of computer security, in particular to a method and system for cross-domain identity association based on multi-relational representation. Background technique [0002] With the rapid development of the Internet, different heterogeneous network identities have been formed due to different organizations in which entities are located and different technical implementation routes in different domains. Due to the difficulty of interoperability and mutual recognition between the two different domains, it has brought many difficulties to the integration and development of cyberspace. Heterogeneous identity alliance is based on alliance chain, smart contract and other technologies. Through the consensus mechanism, the trustworthiness of each system and the identity management service level are agreed, which realizes the unified measurement and evaluation of the credibility of entity identities, and solves the probl...

AI Technical Summary

Problems solved by technology

Although these identity association methods have achieved good results on some data sets, they are not of high practical value in the industry because they only use unilateral entity relationships.

In the real domain, there are usually multiple types of relationships between entities, such as friend relationships, follower relationships, forwarding relationships, etc. Single-type relationship information often has problems such as weak representation capabilities, and cross-domain identity association only uses single-type relationships in some Cannot achieve better results in the scene

Images