A training method and apparatus for a fraud detection model and fraud detection method and apparatus

Abstract

A training method and apparatus for a fraud detection model and a fraud detection method and apparatus are provided. That method and apparatus include obtaining historical operation behavior information of a plurality of sample users in a first historical period and a second historical period, and labeling information of whether a fraud behavior of the sample user occurs in a second historical period; generating a behavior feature vector sequence corresponding to each service scenario for the sample user in a plurality of service scenarios; the behavioral eigenvector sequence is input into theloop neural network to generate the behavioral coding vector. After the behavior coding vectors corresponding to each service scenario are spliced, the classification neural network is inputted to obtain the fraud risk probability. Based on the fraud risk probability and labeling information, the circular neural network and classification neural network are trained to obtain the fraud detection model. The present application can improve the accuracy of judging whether or not the user operation behavior occurring when the user uses the electronic banking is a fraudulent behavior.

Description

technical field [0001] The present application relates to the field of computer information technology, in particular, to a fraud detection model training method and device, and a fraud detection method and device. Background technique [0002] Electronic banking refers to the communication channels that banks open to the public. With the rapid development of the Internet and the popularity of smart terminals, e-banking is a commonly used banking business processing method in e-commerce activities, such as: using e-banking for online transfers, checking account balances, online financial management and other out-of-the-counter services, and Handling business outside the cabinet provides great convenience for users. However, while electronic banking provides convenience for users, there are also many potential safety hazards, such as: account information theft, account information embezzlement, and abnormal operation behaviors such as stolen funds, which damage the interests...

AI Technical Summary

Problems solved by technology

[0004] However, due to the large differences in the operation behaviors of different users, the operation behaviors of the same user in different time periods will also be different. Only based on the current user behavior operation information and machine learning models, the detection and analysis of the current business Whether the user's behavior in the scene is a fraudulent behavior, the accuracy rate is low

Images