Consumption financial product credit scoring method and device based on customer grouping

Abstract

The invention discloses a consumption financial product credit scoring method and device based on customer grouping, which can overcome the technical defects of a traditional single-channel credit scoring model and solve the key problems of a customer grouping credit evaluation method. The method comprises the following steps: obtaining pedestrian credit investigation data of a single customer, and judging one or more customer groups to which the single customer belongs according to a predefined customer grouping rule; and importing the credit investigation data of the customer into the grouping logistic regression score card model of the corresponding customer group, and calculating to obtain a grouping credit score corresponding to each customer group. Based on the thought of customer grouping credit evaluation, the method is applied to risk control work of financial product consumption, so that the defect that an existing credit scoring mode of financial products cannot fully reflect default behavior characteristics of different customer groups is overcome.

Description

Technical field [0001] The present invention relates to the field of financial product wind control technology, in particular to customer credit score technology in consumption finance products. Background technique [0002] With the improvement of economic development and public consumption levels in my country, all kinds of consumption financial products are more popular. The customer's credit score is the main way for customers in the risk control of consumption finance products. Credit score refers to a credit score model using a certain credit score model based on the customer's credit history data. Depending on the customer's credit score, creditors can analyze the consumer's credit rating to give consumers varying degrees of discounts and offers. The most critical is the construction of the credit score model during the process of credit score. The model used to generate credit scores, each model has its own unique rules. [0003] The business of current consumption financ...

AI Technical Summary

Problems solved by technology

The above-mentioned single-channel credit scoring model focuses on reflecting the impact of different traffic channels on credit scoring, but fails to reflect the significant impact of different types of customer groups in the customer base on credit scoring, and has the following disadvantages: (1) The types of customer groups in the channel Many, resulting in low model discrimination

(2) The input variables of different customer groups are quite different, resulting in the variables of the channel model being general input variables among customer groups

(3) Without in-depth portrait distinction, the effect of the model is low, resulting in poor effect of the quota interest rate strategy

[0004] In short, the existing scoring model does not really reflect the impact of customers' personal attributes (such as occupation, age, fixed assets, current credit status) on credit scoring, resulting in a large number of different types of customers adopting a unified model that affects the final evaluation of the model Effect

Images