Risk detection method and device for transaction data and server

Abstract

The invention provides a transaction data risk detection method, a transaction data risk detection device and a server. Based on the method, after a target transaction data group which forms an annular transaction loop and possibly has a preset transaction risk is screened out from a large amount of transaction data, unstructured data and structured data are extracted from the target transaction data group respectively according to a preset extraction rule; calling a preset classification model containing a network model based on a Text CNN-Attention structure to process the two types of data of the unstructured data and the structured data at the same time, so as to obtain a corresponding classification processing result; and determining whether a transaction loop represented by the target transaction data set has a preset transaction risk or not according to a classification processing result. Therefore, whether the transaction loop represented by the target transaction data group really has the preset transaction risk can be accurately determined, the risk detection accuracy of the transaction data is effectively improved, and the detection error is reduced.

Description

technical field [0001] This specification belongs to the field of artificial intelligence technology, and in particular relates to a risk detection method, device and server for transaction data. Background technique [0002] In the transaction data processing scenario, it is often necessary to monitor abnormal capital transaction circuits, so as to supervise and deal with some transaction behaviors with transaction risks in a timely manner. [0003] Only a part of the capital transaction circuits detected based on existing methods may be truly abnormal capital transaction circuits, and there are corresponding transaction risks; while the remaining part belongs to normal capital transaction circuits, and there is no transaction risk. [0004] For example, the originating account A of the transaction data originally planned to send the transaction data to account B, but due to A's personal negligence when initiating the transaction data, the account name of the receiving acco...

AI Technical Summary

Problems solved by technology

[0005] Based on the existing methods, it is often impossible to accurately and effectively identify and determine whether the above-mentioned transaction loops really have corresponding transaction risks, which affects the risk detection of transaction data.

[0006] For the above problems, no effective solutions have been proposed so far

Images