Unlock instant, AI-driven research and patent intelligence for your innovation.

Customer score verification method based on good and bad labels

A verification method and labeling technology, which is applied in the fields of instruments, finance, and data processing applications, can solve problems such as failure to pass cash withdrawal risk control audits, and achieve the effects of shortening timeliness, effective discovery, and reasonable evaluation

Pending Publication Date: 2022-01-11
北京睿知图远科技有限公司
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] For cash withdrawal application, the customer can apply for cash withdrawal to obtain funds after the credit approval of the bank has been approved. The transformation of the customer's cash withdrawal application is accidental. When the customer withdraws the application, it will be reviewed again. After the withdrawal is approved (the withdrawal is passed), the bank will release the money to the customer's bank card. Customers who cannot pass the withdrawal risk control review will be rejected, that is, the withdrawal will be rejected.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Customer score verification method based on good and bad labels
  • Customer score verification method based on good and bad labels
  • Customer score verification method based on good and bad labels

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0081] according to figure 1 As shown, the present invention provides a kind of customer scoring verification method based on good and bad labels, it is characterized in that, comprises:

[0082] Obtain the modeling sample A and the non-modeling sample B, and calculate the scoring index of the loan sample A4 in the modeling sample A; among them,

[0083] The modeling sample A includes a credit application sample A1, a credit approval sample A2, a cash withdrawal application sample A3, and a loan sample A4;

[0084] The non-modeling sample B includes a credit application sample B1, a credit approval sample B2, a cash withdrawal application sample B3, and a loan sample B4;

[0085] Based on the preset scoring variable avoidance principle and superiority principle, construct a good or bad label inference strategy;

[0086] Through the good and bad label inference strategy, the labels of the samples in the modeling sample A and the modeling sample B are respectively identified, ...

Embodiment 2

[0091] according to figure 2 As mentioned above, this technical solution provides an embodiment, the acquisition of modeling sample A and modeling non-sample B includes:

[0092] Obtain the credit application sample A1 within the preset historical time threshold and the target credit application sample B1 within the preset time threshold;

[0093] According to the customer quality information preset by the customer in the authorization application sample A1 and the customer quality information preset by the customer in the target credit application sample B1, calculate the first credit approval rate and the second credit approval rate respectively, and determine the credit approval sample A2 and the credit approval approval rate Sample B2;

[0094] Collect the cash withdrawal impact parameters of customers in the credit approval sample A2 and the credit approval sample B2 respectively, and determine the withdrawal application sample A3 and the withdrawal application sample B...

Embodiment 3

[0102] The technical solution provides an embodiment, the scoring index includes a scoring effect index KS and a scoring variable index IV, wherein the calculation formula of the scoring effect index KS is:

[0103]

[0104] Among them, KS represents the scoring effect index, and F G Cumulative probability distribution function for the estimated scores of well-labeled samples; F B Cumulative probability distribution function for the estimated score of badly labeled samples; Score i It is the i-th score after mixing and sorting good label samples and bad label samples; i=1,2,3...n, where n is the total number of scores.

[0105] The working principle and beneficial effects of the above-mentioned technical scheme are:

[0106] The KS (Kolmogorov-Smirnov) statistic of this technical solution was proposed by two Soviet mathematicians A.N.Kolmogorov and N.V.Smirnov. In statistics, it is a non-parametric test (KS test) based on the cumulative distribution function, which is us...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a customer score verification method based on good and bad labels, and the method comprises the steps of obtaining a modeling sample A and an external modeling sample B, and calculating a score index of a loan sample A4 in the modeling sample A, wherein the modeling sample A comprises a credit granting application sample A1, a credit granting passing sample A2, a cash withdrawal application sample A3 and a loan sample A4; on the basis of a preset scoring variable evasion and better principle, building a good and bad label inference strategy, and through the good and bad label inference strategy, performing label determination on samples in the modeling sample A and the modeling external sample B, and determining corresponding label samples, wherein the modeling external sample B comprises a credit granting application sample B1, a credit granting passing sample B2, a cash withdrawal application sample B3 and a loan sample B4; and calculating a scoring index corresponding to the label sample, drawing a monitoring report according to the scoring index, analyzing the monitoring report, and performing scoring verification on the scoring index of the modeling external sample B.

Description

technical field [0001] The invention relates to the technical field of customer rating verification, in particular to a customer rating verification method based on good and bad labels. Background technique [0002] The loan application scoring model is an important technology in the field of commercial bank loan risk control. It uses statistical analysis and machine learning algorithms to legally collect customer authorization information, establishes scores to comprehensively evaluate customer credit, gives credit approval decisions, and uses scores for credit approval. , which is usually more efficient, lower cost, and more objective than manual approval, and there will be no situation where different approvers give different decisions. [0003] After the development of the application score is completed and deployed online, and to participate in risk decision-making, it is necessary to continuously track and monitor the use effect of the application score. Through monito...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06Q40/02G06Q10/06
CPCG06Q10/063114G06Q10/0635G06Q10/06393G06Q40/03
Inventor 肖玉龙
Owner 北京睿知图远科技有限公司