Customer layering method and device

Abstract

The invention aims to provide a customer layering method and device. The method comprises: determining hierarchical evaluation of a user in a plurality of cycles based on historical performance data of the user to be evaluated and pre-stored hierarchical boundary information; based on the hierarchical evaluation of the user and other multiple users in multiple periods, calculating hierarchical mobility of multiple hierarchies; and if the hierarchical mobility does not meet the predetermined condition, continuously increasing the boundary value range of each layer of the current hierarchical boundary information to recalculate the hierarchical mobility until the calculated hierarchical mobility meets the predetermined condition. The embodiment of the invention has the following advantages: when the customer hierarchy of the user is predicted, the customer hierarchy of the user in a plurality of periods is obtained, and the boundary of the customer hierarchy is continuously widened through an iterative adjustment feedback mode, so that the finally obtained customer hierarchy is relatively stable, and then the serious deviation of the risk estimation of the customer when the customer state is greatly changed is avoided.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a method and device for client stratification. Background technique [0002] Based on the existing technology, after the customer registers and applies for the first time on the platform that provides credit business, the platform distributes the customer to the undertaking institution to complete the initial matching of capital assets. After a customer completes a loan or has a loan on the platform for a certain period of time, he becomes an old loan customer. At this time, the platform will regularly re-evaluate the customer's risk based on the customer's historical loan and repayment information and external credit information. Make appropriate adjustments to customer risks to provide services that better match risks. Customer risks are deteriorating rapidly, and timely measures must be taken to avoid capital losses. In addition, whether from the perspective of customers or...

AI Technical Summary

Problems solved by technology

[0003] However, based on this method, the customer stratification will change over time, and the stability of the stratification is poor. The main reasons include: 1) The risk model score in the loan is used to formulate the stratification; The design goal does not completely match the stratification goal. The goal of the loan risk model is to pursue the risk position of the customer among all customer groups at a certain point in time, while the stratification requires a certain period of time while completing the division of risk levels. Continuity; 2) Changes in the state of the customer lead to model prediction risk deviation; the model generally predicts risk from a group of customers in a similar state, combined with variables such as customer behavior, and the difference in the distribution of group characteristics should not be too large. However, there is a change in the customer status of the loan customer, such as changing from an active customer to an inactive customer, which may cause a serious deviation in the risk estimation of the customer

Images