Fraud node identification method based on graph embedded representation and recurrent neural network

Abstract

The invention relates to a fraud node identification method based on graph embedding representation and a recurrent neural network. The fraud node identification method comprises the following steps:1) obtaining a data set containing a relational network and node behaviors as an original data set; 2) preprocessing the original data set to obtain graph structure data and node labels; 3) generatingnode embedding representation by using the graph structure data at different time points; 4) inputting the node embedding representations of the same node at different time points into the recurrentneural network according to a time sequence to obtain a final node embedding representation; and 5) training the neural network model by using the final node embedded representation, and predicting the fraud risk of the user by using the trained neural network model. Compared with the prior art, the method has the advantages that features can be directly extracted from a user relationship networkstructure, user node features are not needed, dependence on external features is not needed, and the method adapts to a real environment dynamic graph.

Description

technical field [0001] The invention relates to the technical field of network security, in particular to a fraudulent node identification method based on graph embedding representation and cyclic neural network. Background technique [0002] With the advancement of information technology, the complex structure of semi-structured data such as social networks, paper networks, and webpage networks has brought great challenges to traditional graph data processing algorithms. At the same time, due to the huge volume of these network data , simply applying the neural network model for training will cause serious performance problems. In addition, the data in real life is often updated dynamically. The neighbor nodes of nodes and the structure of the entire network will change over time. How to deal with nodes The dynamic change in the time dimension is also a problem worthy of attention. Contents of the invention [0003] The object of the present invention is to provide a met...

AI Technical Summary

Problems solved by technology

[0002] With the advancement of information technology, the complex structure of semi-structured data such as social networks, paper networks, and webpage networks has brought great challenges to traditional graph data processing algorithms. At the same time, due to the huge volume of these network data , simply applying the neural network model for training will cause serious performance problems. In addition, the data in real life is often updated dynamically. The neighbor nodes of nodes and the structure of the entire network will change over time. How to deal with nodes The dynamic change in the time dimension is also a problem worthy of attention

Images