Fraud user identification method and device, computer equipment and storage medium

Abstract

The invention discloses a fraud user identification method and device, computer equipment and a storage medium. The method comprises the following steps: carrying out data cleaning on obtained nodes corresponding to claim settlement data to obtain cleaned nodes; Dividing the cleaned nodes into a plurality of subgraphs in parallel through spectral clustering; Respectively clustering the plurality of subgraphs to obtain a network community comprising a plurality of clustering clusters; Obtaining a target node corresponding to a high-risk user tag in the network community through tag propagationaccording to the node tag initially set in the network community; And if the feature vector of the network association has the feature vector which is the same as the target feature vector corresponding to the target node, obtaining the corresponding network association and carrying out identification of the fraud association. According to the method, the network is cut through the clustering algorithm, the network scale is reduced, the network structure is optimized, the risk identification accuracy is improved, and fraudulent users and communities are accurately positioned.

Description

technical field [0001] The present invention relates to the technical field of fraudulent user identification, in particular to a fraudulent user identification method, device, computer equipment and storage medium. Background technique [0002] At present, in the process of claim settlement in the insurance industry, social analysis has begun to be applied to the business scenario of user risk analysis. Most of the data is manually entered, there are abnormal data, and the business logic is relatively complicated. The analysis of a single algorithm obviously has the following shortcomings: [0003] 1) There is a large-scale network, and the risk score is abnormally high due to the large number of matching rules in the network; [0004] 2) Lack of a risk user energy propagation model, which cannot cover risk cohesion analysis; [0005] 3) Lack of global consideration, artificially defined features do not contain network structure information. Contents of the invention ...

AI Technical Summary

Problems solved by technology

[0006] Embodiments of the present invention provide a fraudulent user identification method, device, computer equipment, and storage medium, aiming to solve the problem of using a single analysis algorithm in the business scenario where social analysis is applied to user risk analysis in the prior art due to the existence of a large-scale network. The risk assessment results are inaccurate, and due to the lack of risk user energy transmission, the problem of risk cohesion analysis cannot be covered

Images