Methods, systems, and computer program products for estimating accuracy of linking of customer relationships

• US 8,170,998 B2
• Filed: 09/12/2007
• Issued: 05/01/2012
• Est. Priority Date: 09/12/2007
• Status: Expired due to Fees

• Current Assignee: Liberty Peak Ventures, LLC (Dominion Harbor Enterprises, LLC)
• Original Assignee: American Express Travel Related Services Company, Inc. (American Express Company)
• Inventors: Weissman, Mary, TRESCH, Sara, FITZGERALD, Gary Stephen, CHURI, Prashant Sharad
• Primary Examiner(s): Lu, Kuen
• Assistant Examiner(s): Rayyan, Susan F
• Application Number:
  US11/854,259
• Publication Number:
  US 20090070289A1
• Time in Patent Office: 1,693 Days
• Field of Search: 707200-206, 707/610, 707/690, 707/692
• US Class Current: 707/690
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 40/02 Banking, e.g. interest calculation or account maintenance credit or loans G06Q40/03

Abstract

The disclosed methods, systems, and computer-program products allow a business to estimate linking errors in customer relationships in a database and to identify metrics that improve the linking accuracy. In an embodiment, a plurality of sample customer relationships are selected from a database to form a sample data set that is statistically representative of the database. Potential linking errors are then identified within the sample data set. The identified potential linking errors are then validated to identify actual linking errors in the sample data set. Once validated, the actual linking errors within the sample data set are used to estimate linking errors within the database. Further, the estimated linking errors in the database may be analyzed to identify one or more factors that contribute to the linking errors.

Claims

1. A method comprising:

selecting, by a computer based system for estimating accuracy of customer relationships, from a database comprising customer relationships data a limited stratified subset of customer relationships data to form a sample data set;

identifying, by the computer based system, linked relationships associated with different customers and unlinked relationships associated with a same customer as potential linking errors in the sample data set;

validating, by the computer based system, the potential linking errors to identify actual linking errors of the sample data set;

storing, by the computer based system, an error type designation of the actual linking errors of the sample data set; and

estimating, by the computer based system, the linking errors within the database by extrapolating the identified actual linking errors of the sample data set to estimate the linking errors within a data set of the database larger than the sample data set, wherein the estimate includes the linking error type designation.

2. The method of claim 1, wherein the selecting step comprises selecting from the database a plurality of sample customer relationships, wherein each relationship includes at least one of (i) a name, (ii) a street address, (iii) a city, (iv) a state, (v) a zip code, (vi) a social security number, (vii) a date of birth, (viii) a supplemental social security number, (ix) an relationship number, (x) a telephone number, and/or (xi) a customer number associated with the respective customer relationship.

3. The method of claim 1, wherein the identifying step comprises:

selecting a plurality of pairs of linked relationships from the database to form the sample data set;

comparing, for each selected pair of linked relationships, elements of data associated with each relationship within the selected pair of linked relationships to generate a set of elemental scores;

combining, for each selected pair of linked relationships, the set of elemental scores to generate a composite score; and

analyzing the composite score of each selected pair of linked relationships to identify a potential linking error associated with the selected pair of linked relationships.

4. The method of claim 3, wherein the comparing step comprises applying a rule to each element to determine an elemental score;

and wherein the combining step comprises applying a rule to generate the composite score from the set of elemental scores.

5. The method of claim 1, wherein the identifying step comprises:

selecting from the database a plurality of relationships linked to customers to form the sample data set;

retrieving, for each customer in the sample data set, a set of candidate relationships that are potentially linked to the respective customer;

comparing, for each customer in the sample data set, elements of data associated with (i) an relationship linked to the respective customer and (ii) each candidate relationship potentially linked to the respective customer to generate a set of elemental scores for each candidate relationship;

combining, for each candidate relationship, the set of elemental scores to generate a composite score; and

analyzing the composite score of each candidate relationship to identify a potential linking error associated with the candidate relationship.

6. The method of claim 5, wherein the retrieving step comprises applying a key based on a portion of at least one element of data associated with the respective customer to the plurality of relationships within the sample to identify the set of candidate relationships.

7. The method of claim 5, wherein the comparing step comprises applying a rule to each element to determine an elemental score;

and wherein the combining step comprises applying a rule to generate the composite score from the set of elemental scores.

8. The method of claim 1, further comprising:

analyzing, by the computer based system, the estimated linking errors to identify at least one factor contributing to the linking errors; and

modifying, by the computer based system, a linking procedure in response to the at least one identified factor.

9. An article of manufacture including a non-transitory, tangible computer readable medium having instructions stored thereon that, in response to execution by a computer-based system for estimating the accuracy of the customer relationships, cause the computer-based system to perform operations comprising:

selecting, by the computer based system, from a database comprising customer relationships data a limited stratified subset of customer relationships data to form a sample data set;

identifying, by the computer based system, linked relationships associated with different customers and unlinked relationships associated with a same customer as potential linking errors in the sample data set;

validating, by the computer based system, the potential linking errors to identify actual linking errors of the sample data set;

storing, by the computer based system, an error type designation of the actual linking errors of the sample data set; and

estimating, by the computer based system, the linking errors within the database by extrapolating the identified actual linking errors of the sample data set to estimate the linking errors within a data set of the database larger than the sample data set, wherein the estimate includes the linking error type designation.

10. The non-transitory computer-readable medium of claim 9, wherein each sample customer relationship includes at least one of (i) a name, (ii) a street address, (iii) a city, (iv) a state, (v) a zip code, (vi) a social security number, (vii) a date of birth, (viii) a supplemental social security number, (ix) an relationship number, (x) a telephone number, and/or (xi) a customer number associated with the respective customer relationship.

11. The non-transitory computer-readable medium of claim 9, wherein the identifying comprises:

selecting a plurality of pairs of linked relationships from the database to form the sample data set;

comparing, for each selected pair of linked relationships, elements of data associated with each relationship within the selected pair of linked relationships to generate a set of elemental scores;

combining, for each selected pair of linked relationships, the set of elemental scores to generate a composite score; and

analyzing the composite score of selected each pair of linked relationships to identify a potential linking error associated with the selected pair of linked relationships.

12. The non-transitory computer-readable medium of claim 11, wherein the comparing comprises applying a rule to each element to determine an elemental score;

and wherein the combining comprises applying a rule to generate the composite score from the set of elemental scores.

13. The non-transitory computer-readable medium of claim 9, wherein the identifying comprises:

selecting from the database a plurality of relationships linked to customers to form the sample data set;

retrieving, for each customer in the sample data set, a set of candidate relationships that are potentially linked to the respective customer;

comparing, for each customer in the sample data set, elements of obtained data associated with (i) an relationship linked to the respective customer and (ii) each candidate relationship potentially linked to the respective customer to generate a set of elemental scores for each candidate relationship;

combining, for each candidate relationship, the set of elemental scores to generate a composite score; and

analyzing the composite score of each candidate relationship to identify a potential linking error associated with the candidate relationship.

14. The non-transitory computer-readable medium of claim 13, wherein the retrieving comprises applying a key based on a portion of at least one element of data associated with the respective customer to the plurality of relationships within the sample to identify the set of candidate relationships.

15. The non-transitory computer-readable medium of claim 13, wherein the comparing comprises applying a rule to each element to determine an elemental score and wherein the combining comprises applying a rule to generate the composite score from the set of elemental scores.

16. The non-transitory computer-readable medium of claim 9, further comprising:

analyzing the estimated linking errors to identify at least one factor contributing to the linking errors; and

modifying a linking procedure in response to the at least one identified factor.

17. A system comprising:

a tangible, non-transitory memory communicating with a processor for estimating the accuracy of the customer relationships,

the tangible, non-transitory memory having instructions stored thereon that, in response to execution by the processor, cause the processor to perform operations comprising;

select, by the processor, from a database comprising customer relationships data a limited stratified subset of customer relationships data to form a sample data set;

identify by the processor, linked relationships associated with different customers and unlinked relationships associated with a same customer as potential linking errors in the sample data set;

validate by the processor, the potential linking errors to identify actual linking errors of the sample data set;

store, by the processor, an error type designation of the actual linking errors of the sample data set; and

estimate by the processor, the linking errors within the database by extrapolating the identified actual linking errors of the sample data set to estimate the linking errors within a data set of the database larger than the sample data set, wherein the estimate includes the linking error type designation.

18. The computer based system of claim 17, wherein the processor is further configured to:

select a plurality of pairs of linked relationships from the database to form the sample data set;

compare, for each selected pair of linked relationships, elements of obtained data associated with each relationship within the selected pair of linked relationships to generate a set of elemental scores;

combine, for each selected pair of linked relationships, the set of elemental scores to generate a composite score; and

analyze the composite score of each selected pair of linked relationships to identify a potential linking error associated with the selected pair of linked relationships.

19. The computer based system of claim 17, wherein the processor is further configured to:

select from the database a plurality of relationships linked to customers to form the sample data set;

retrieve, for each customer in the sample data set, a set of candidate relationships that are potentially linked to the respective customer;

compare, for each customer in the sample data set, elements of data associated with (i) an relationship linked to the respective customer and (ii) each candidate relationship potentially linked to the respective customer to generate a set of elemental scores for each candidate relationship;

combine, for each candidate relationship, the set of elemental scores to generate a composite score; and

analyze the composite score of each candidate relationship to identify a potential linking error associated with the candidate relationship.