Recommendation method and system with score reduction mechanism based on graph neural network

Abstract

The invention relates to a graph neural network-based recommendation method and system with a score reduction mechanism, and the method comprises the steps: obtaining a user data set and user features, and carrying out the score initialization of user nodes; performing positive scoring and negative scoring on the user to obtain a score adding vector and a score reducing vector, and evaluating the safety of the user according to the score reducing vector; when a friend is recommended to a certain user, summing each component of the score adding vector through the weight, and subtracting the sum of each component in the score subtraction vector through the weight; and updating the score addition vector and the score reduction vector of the user at set intervals, and re-detecting the security of the user. According to the method, the problem that the number of label nodes in graph data is small can be effectively solved, so that the abnormal account in the recommendation system can be detected more quickly and accurately.

Description

technical field [0001] The invention relates to the technical field of recommendation systems, in particular to a recommendation method and system based on a graph neural network with a score reduction mechanism. Background technique [0002] Under the premise that the recommendation system is widely used, there are many situations in which the actual users of social accounts have changed, or the social accounts have been stolen. Under this premise, the structure of the corresponding recommendation system will also change, so the friend recommendation system should also generate corresponding At the same time, an early warning should be given to social accounts that have been hacked or have been suddenly and widely blocked. [0003] like figure 1 As shown, the prior art proposes a method and system for detecting abnormal accounts in a recommendation system, which includes the following steps: Step S101 : constructing a social network account graph; wherein the account graph...

AI Technical Summary

Problems solved by technology

[0007] The main purpose of the present invention is to propose a recommendation method and system based on a graph neural network with a score reduction mechanism, which solves the problem of changes in the recommendation system caused by changes in the status or relationship of friends in the friend system of the recommendation system, and realizes the recommendation of friends. Dynamic assessment and early warning of accounts that may have problems

Images