Aggregated merchant model construction and abnormal risk aggregated merchant monitoring method

Abstract

The invention discloses an aggregated merchant model construction and abnormal risk aggregated merchant monitoring method. The method comprises the steps of firstly extracting an operation range and an aggregated merchant address of an aggregated merchant through an nlp technology, correcting the operation range and the address of the commercial tenant, and characterizing the corrected operation range and the address of the merchant by using a trained word vector, and generating an aggregated merchant characterization vector. The method has the advantages that the address correction technologyand the merchant operation range correction technology are added, the situation that merchant addresses and aggregated merchants lack the operation range or information is incomplete is effectively filled up, and the accuracy of abnormal merchant recognition is improved; aggregated merchants with similar addresses and similar operation ranges can be effectively clustered together, so that the accuracy of abnormal merchant identification is improved; and finally, aiming at the condition that the proportion of abnormal merchants is very low in an actual scene, an algorithm training model with supervised anomaly detection is innovatively applied, so that the accuracy of problem abnormal merchant identification is improved.

Description

technical field [0001] The invention relates to the field of electronic information technology, in particular to a method for constructing an aggregation merchant model and monitoring abnormal risk aggregation merchants. Background technique [0002] At present, aggregate payment is developing rapidly and accounts for a large proportion in actual transactions. However, many of the aggregated merchants are merchants not in their own system, and they can obtain less information from their partners (only some information such as transaction amount, time, type, encrypted user id, etc.), and have a large compliance risk exposure. In actual monitoring, there is a problem of difficult monitoring. At present, there are few relevant methods for monitoring this risk. At the same time, only through some transaction characteristics of the store to monitor, the coverage rate of black merchants is low, and the false arrest rate of normal merchants is high. Through the detailed analysis...

AI Technical Summary

Problems solved by technology

At the same time, the proportion of abnormal risk merchants in our total number of merchants is relatively low

Images