Lending fraud detection model training method, lending fraud detection method and device

Abstract

The invention provides a lending fraud detection model training method, a lending fraud detection method and a lending fraud detection device. The lending fraud detection model training method comprises the following steps: obtaining identity information of a plurality of sample users, user bank running information and fraud labeling information corresponding to each sample user; constructing identity feature vector vased on identity information, and constructing a pipeline characteristic vector according to the pipeline information of the user bank; splicing the identity feature vector and pipeline feature vector to generate the target feature vector which is used to characterize the status of each sample user and expenditure revenue. The target feature vector is inputted to the target neural network, and the fraud detection result of the target feature vector is obtained. According to the fraud detection results and the corresponding fraud labeling information, the target neural network is trained to obtain the lending fraud detection model. The application can improve the identification efficiency and the identification accuracy of the fraudulent user by the credit platform, andgreatly save the human cost.

Description

technical field [0001] The present application relates to the technical field of machine learning, in particular, to a loan fraud detection model training method, a loan fraud detection method and a device. Background technique [0002] With the rapid development of Internet finance, the incidence of Internet fraud and credit risk has increased. According to statistics, in recent years, the non-performing asset rate of the consumer finance industry in China has been on the rise. China's Internet fraud risk has ranked among the top three in the world. Cybercrime brings economic losses of up to 445 billion US dollars to the world every year. Penetration of different industries. [0003] Internet financial risks include credit risk and operational risk. Credit risk means that customers have no intention of repaying the loan when they initiate a loan request. Among them, fraud is the highest credit risk, and more than 50% of the bad debt losses of consumer finance come from fr...

AI Technical Summary

Problems solved by technology

[0006] It is precisely because the credit review of borrowing users on the credit platform basically relies on the expert experience of business personnel to make judgments. Since the bank's historical flow information is often complicated and the amount of information involved is relatively large, if it is purely manually reviewed Not only is it inefficient, but it is also difficult to draw an overall correct conclusion by integrating bank flow data and user personal information

Images