SYSTEM, METHOD, AND NON-TRANSITORY COMPUTER-READABLE STORAGE MEDIA FOR PREDICTING A CUSTOMER'S CREDIT SCORE

• US 20160225073A1
• Filed: 01/30/2015
• Published: 08/04/2016
• Est. Priority Date: 01/30/2015
• Status: Abandoned Application

• Current Assignee: Walmart Apollo, LLC (WalMart Inc.)
• Original Assignee: Wal-Mart Stores Texas LLC (WalMart Inc.)
• Inventors: Bash, David, Pavlidis, Yannis, Xiao, Fei
• Application Number:
  US14/610,405
• Publication Number:
  US 20160225073A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06Q 40/03 Credit; Loans; Processing thereof

Abstract

In different embodiments of the present invention, systems, methods, and computer-readable storage media establishes a predicted credit score for a target customer that does not have an established credit history and/or a FICO credit score is provided. An indicator of the purchasing power of the target customer is established as a function of stored purchasing data. A machine learning model is used to establish a predicted credit score associated with the target customer as a function of the purchasing power indicator of the target customer and the machine learning model.

Claims

1. A system for establishing a predicted credit score for a target customer that does not have an established credit history and/or a FICO credit score, comprising:

a memory unit configured to store purchasing data associated with the target customer, the purchasing data including the target customer'"'"'s purchasing transactions with a retailer, the purchasing data for the target customer being stored in an associated customer account;

a customer purchasing power indicator unit coupled to the memory unit and being configured to establish an indicator of the purchasing power of the target customer as a function of the purchasing data in the associated customer account and to store the established purchasing power indicator of the target customer in the memory unit, the memory unit being further configured to store a purchasing power indicator and a credit score indicator for each of a plurality of other customers;

a machine learning model unit coupled to the memory unit and being configured to train a machine learning model as a function of the purchasing power indicator and the credit score indicator associated with each of the plurality of other customers; and

,

a credit score prediction unit coupled to the memory unit and the customer purchasing power indicator unit and being configured to establish a predicted credit score associated with the target customer as a function of the purchasing power indicator of the target customer and the machine learning model and to store the predicted credit score in the memory unit.

2. A system, as set forth in claim 1, the memory unit is further configured to store purchasing data associated with each of the plurality of other customers.

3. A system, as set forth in claim 2, wherein the customer purchasing indicator unit is further configured to establish the purchasing power indicator for each of the plurality of other customers as a function of the purchasing data associated the respective one of the plurality of other customers.

4. A system, as set forth in claim 1, wherein the credit score indicator of each of the plurality of other customers is a function of an actual FICO score of the respective customer.

5. A system, as set forth in claim 4, wherein the credit score indicator is a normalized value.

6. A system, as set forth in claim 1, wherein the credit score indicator of each of the plurality of other customers is an estimated credit score.

7. A system, as set forth in claim 6, wherein the credit score indicator of each of the plurality of other customers is established using:

T*cⁿ,

where T is the respective customer'"'"'s maximum monthly total credit card spending at the retailer, c is a predetermined constant, and n is a number of credit cards used by the respective customer at the retailer.

8. A system, as set forth in claim 6, wherein the credit score indicator of each of the plurality of other customers is established using:

a1*M+a2*S+a3*H+a4*D+a5*n,

where, M is the maximum credit usage per month among all months, S is the past year average monthly spending with credit cards, H is the overall credit card spending history length in number of months, D is the standard deviation of the past year monthly credit cards spending, n is the total number of credit cards used, and a1-a5 are predetermined constants.

9. A system, as set forth in claim 1, wherein the purchasing data associated with the target customer includes at least one of the following:

average monthly spending at the retailer, average spending per visit at the retailer, spending history at the retailer, average number of category of goods shopped at the retailer, standard deviation of monthly spending at the retailer, and average monthly per category of goods spending.

10. A system, as set forth in claim 1, wherein the machine learning model is one of a logistic regression model, a linear regression model, a smoothing splines model, a generalized additive model, and a regression tree model.

11. A system, as set forth in claim 1, wherein the machine learning model unit is further configured to be adapt the machine learning model as a function of past decisions.

12. A method for establishing a predicted credit score for a target customer that does not have an established credit history and/or a FICO credit score, including the steps of:

storing, on a memory unit, purchasing data associated with the target customer, the purchasing data including the target customer'"'"'s purchasing transactions with a retailer, the purchasing data for the target customer being stored in an associated customer account;

establishing, using a customer purchasing power indicator unit couple to the memory unit, an indicator of the purchasing power of the target customer as a function of the purchasing data in the associated customer account and storing the established purchasing power indicator of the target customer in the memory unit, the memory unit being further configured to store a purchasing power indicator and a credit score indicator for each of a plurality of other customers;

training, using a machine learning model unit coupled to the memory unit, a machine learning model as a function of the purchasing power indicator and the credit score indicator associated with each of the plurality of other customers; and

,

establishing, using a credit score prediction unit coupled to the memory unit and the customer purchasing power indicator unit, a predicted credit score associated with the target customer as a function of the purchasing power indicator of the target customer and the machine learning model and storing the predicted credit score in the memory unit.

13. A method, as set forth in claim 12, including the step of storing purchasing data associated with each of the plurality of other customers in the memory unit.

14. A method, as set forth in claim 13, including the step of establishing the purchasing power indicator for each of the plurality of other customers as a function of the purchasing data associated the respective one of the plurality of other customers.

15. A method, as set forth in claim 12, wherein the credit score indicator of each of the plurality of other customers is a function of an actual FICO score of the respective customer.

16. A method, as set forth in claim 15, wherein the credit score indicator is a normalized value.

17. A method, as set forth in claim 12, wherein the credit score indicator of each of the plurality of other customers is an estimated credit score.

18. A method, as set forth in claim 17, wherein the credit score indicator of each of the plurality of other customers is established using:

T*cⁿ,

where T is the respective customer'"'"'s maximum monthly total credit card spending at the retailer, c is a predetermined constant, and n is a number of credit cards used by the respective customer at the retailer.

19. A method, as set forth in claim 17, wherein the credit score indicator of each of the plurality of other customers is established using:

a1*M+a2*S+a3*H+a4*D+a5*n,

where, M is the maximum credit usage per month among all months, S is the past year average monthly spending with credit cards, H is the overall credit card spending history length in number of months, D is the standard deviation of the past year monthly credit cards spending, n is the total number of credit cards used, and a1-a5 are predetermined constants.

20. A method, as set forth in claim 12, wherein the purchasing data associated with the target customer includes at least one of the following:

average monthly spending at the retailer, average spending per visit at the retailer, spending history at the retailer, average number of category of goods shopped at the retailer, standard deviation of monthly spending at the retailer, and average monthly per category of goods spending.

21. A method, as set forth in claim 12, wherein the machine learning model is one of a logistic regression model, a linear regression model, a smoothing splines model, a generalized additive model, and a regression tree model.

22. A method, as set forth in claim 12, wherein the machine learning model unit is further configured to be adapt the machine learning model as a function of past decisions.

23. One or more non-transitory computer-readable storage media, having computer-executable instructions embodied thereon, wherein when executed by at least one processor, the computer-executable instructions cause the processor to operate as a:

a customer purchasing power indicator unit coupled to a memory unit and being configured to establish an indicator of the purchasing power of the target customer as a function of the purchasing data in the associated customer account and to store the established purchasing power indicator of the target customer in the memory unit, the memory unit being further configured to store a purchasing power indicator and a credit score indicator for each of a plurality of other customers;

a machine learning model unit coupled to the memory unit and being configured to train a machine learning model as a function of the purchasing power indicator and the credit score indicator associated with each of the plurality of other customers; and

,

a credit score prediction unit coupled to the memory unit and the customer purchasing power indicator unit and being configured to establish a predicted credit score associated with the target customer as a function of the purchasing power indicator of the target customer and the machine learning model and to store the predicted credit score in the memory unit.