Method for training anti-fraud model, method for identifying financial fraud behavior and related products thereof

Abstract

The invention relates to a method and electronic equipment for training an anti-fraud model for identifying financial fraud behaviors, and a method, electronic equipment and a system for identifying financial fraud behaviors. The method for training the anti-fraud model is executed by a processor, and comprises the following steps: acquiring sample features about financial fraud behaviors as training data; inputting the training data into the anti-fraud model to perform clustering training on the anti-fraud model; in response to outputting the target class cluster by the anti-fraud model, completing clustering training of the anti-fraud model, wherein the cluster center point in the target class cluster represents a normal financial transaction behavior, and the inter-class-cluster distance of each class cluster in the target class cluster and the inter-class-cluster distance between the elements in each class cluster meet a predetermined evaluation index. Through the technical scheme of the invention, unsupervised training of the anti-fraud model can be realized without depending on a sample label, so that a powerful guarantee is provided for accurate financial fraud behavior identification by using the anti-fraud model.

Description

technical field [0001] The present invention generally relates to the technical field of financial risk control. More specifically, the present invention relates to a method and electronic device for training an anti-fraud model for identifying financial fraud, a method, electronic device and system for identifying financial fraud. Background technique [0002] This section is intended to provide a background or context for implementations of the invention that are recited in the claims. The descriptions herein may include concepts that could be explored, but not necessarily concepts that have been previously thought of or explored. Therefore, unless otherwise indicated herein, what is described in this section is not prior art to the description and claims in this application and is not admitted to be prior art by inclusion in this section. [0003] With the popularization of Internet technology, Internet-based electronic banking has been further developed, and it has bec...

AI Technical Summary

Problems solved by technology

However, existing machine learning models have relatively high quality requirements for sample data, which rely on labeled sample data

However, in the actual model training process, especially for fraudulent behavior model training, the proportion of effective sample data is small and the timeliness of labels is poor, which makes the actual effect of the trained model poor and cannot meet actual needs.

Images