An abnormal transaction detection model training method and device and an abnormal transaction detection method and device

Abstract

The invention provides an abnormal transaction detection model training method and device and an abnormal transaction detection method and device. The training method of the abnormal transaction detection model comprises the steps that an obtained first transaction sample data set is received, and all first transaction sample data in the first transaction sample data set are normal transaction sample data; feature extraction processing is performed on the first transaction sample data set to obtain a first transaction sample feature set; On the basis of the first transaction sample feature set, an unsupervised machine learning algorithm is adopted for training to obtain an abnormal transaction detection model, and the core position and the radius of each cluster obtained on the basis of the unsupervised machine learning algorithm are recorded.

Description

technical field [0001] The present invention relates to the intersecting field of machine learning and financial transactions, in particular to a training method and device for an abnormal transaction detection model, a method and a device for detecting an abnormal transaction, a computing device and a computer-readable storage medium storing computer programs. Background technique [0002] With the prosperity and development of science and technology and social economy, financial transactions (for example, Internet-based financial transactions, etc.) are becoming more frequent and more important. For example, financial institutions such as banks will evaluate financing credit lines and issue loans based on the financial transactions of enterprises. [0003] However, due to the low cost and large profit potential of fraud in financial transactions, there will be enterprises defrauding profits by constructing false financial transactions, for example, fraudulently obtaining l...

AI Technical Summary

Problems solved by technology

[0003] However, due to the low cost and large profit potential of fraud in financial transactions, there will be enterprises defrauding benefits by constructing false financial transactions, for example, fraudulently obtaining loans from banks, etc.

In the face of abnormal transactions (for example, counterfeit transactions), the traditional method of checking forgery through invoices and other means is not real-time, and cannot meet the efficient demands of banks and enterprises in the Internet era. However, if the supervised machine learning method When performing real-time monitoring, a large number of labeled samples are required for training, and the collection and labeling of samples will also consume a lot of time and labor costs, and there are cases of misjudgment

Images