An unusual transaction identification method, device and storage medium based on abnormal financial features

Abstract

The invention provides an unusual transaction identification method, device and storage medium based on abnormal financial features. According to the invention, user-defined transaction entropy features, transaction activeness and amount statistical features are extracted by using a large amount of to-be-detected financial transaction flow information data suspected to be abnormal or related to certain abnormal accounts, and abnormal financial feature vector representation is constructed, and then based on the abnormal feature vectors, classification identification is carries out on the to-be-detected transaction account through a voting classifier to determine whether the to-be-detected transaction account is in a pyramid-selling abnormal transaction relation. The suspicious financial transaction relationship identification method proposed by the invention realizes the relatively rich abstraction and expression of the rule characteristics of complex behaviors of the transaction subject by using the financial transaction flow data, and obtains a relatively good detection and identification effect of the abnormal financial transaction relationship. The abnormal financial transactionrelationship detection result provided by the method can assist the relevant staff to conduct investigation and judgment on abnormal financial activities such as pyramid sales, which can improve workefficiency and improve the accuracy of research and judgment to a certain extent.

Description

technical field [0001] The present invention relates to the field of financial transactions, in particular to an abnormal transaction identification method based on heterogeneous financial features, equipment and a storage medium. Background technique [0002] The financial system is one of the important pillars of modern economic development. With the development of the network and information technology of the financial system, the capital flow is accelerating, and the daily transaction flow of financial transaction institutions such as banks can reach millions or even tens of millions. , which provides great convenience for economic criminal activities hidden in the financial network. In the massive transaction data, there are more complex customer transaction behavior patterns and rules and other in-depth information, which can be used to identify and detect abnormal transaction behavior hidden in the data. Abnormal financial transactions include financial transactions ...

AI Technical Summary

Problems solved by technology

However, this method relies on the existence of relatively complete networks or local topological structures with a large amount of information in the data. In the case of incomplete related structures, the recognition rate or accuracy rate will be low.

The application of machine learning methods in the field of identification of abnormal financial transactions in pyramid schemes is still relatively small. Existing mathematical models such as branch trees still rely too much on the complex patterns and evolution laws of the development of pyramid schemes extracted manually, and have not fully explored the More intelligent and efficient statistical methods, and there are certain defects in the processing of massive data

At present, there is no effective machine learning solution based on heterogeneous financial characteristics, which can effectively detect and identify abnormal financial transaction relationships in pyramid schemes

Images