Method and system for approximate matching of data records

• US 20030195836A1
• Filed: 04/09/2003
• Published: 10/16/2003
• Est. Priority Date: 12/18/2000
• Status: Active Grant

• Current Assignee: Powerloom
• Original Assignee: Powerloom
• Inventors: Mani, Ganesh, Fink, Eugene, Hayes, Philip J.
• Granted Patent: US 7,801,878 B2
• Time in Patent Office: Days
• US Class Current: 705/37
• CPC Class Codes: G06Q 20/102 Bill distribution or payments
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes
• CPC Class Codes: G06Q 40/04 Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange

Abstract

The present invention is directed to a method of matching a data record with a reference record. The method comprises receiving the data record; searching stored reference records for a reference record that is an exact match for the received data record; and if an exact matching reference record is not found, calculating a distance between the received data record and at least one stored reference record, and determining the at least one stored reference record to be an approximate match if the calculated distance is less than a pre-determined threshold.

The inventive method and system can be used in a number of applications, including invoice processing in an Accounts Payable system, record conciliation in a securities trading system, finding preferred bonds in a bond trading system, etc.

Approximate matching records are ranked, and the closest match, or highest ranking match, may be automatically processed if within a predetermined threshold.

Claims

1. A method of matching a data record with a reference record in a system, comprising:

(A) receiving the data record;

(B) calculating a distance between the received data record and a reference record; and

(C) determining that the reference record approximately matches the received data record if the calculated distance is less than a pre-determined threshold.

2. The method of claim 1, step (C) additionally comprises:

determining that the reference record exactly matches the received data record if the calculated distance is zero.

3. The method of claim 2, repeating steps (B) and (C) until a reference record is determined to exactly match the received data record.

4. The method of claim 2, repeating steps (B) and (C) until a reference record is determined to exactly match the received data record, or until a predetermined maximum number of distances have been calculated.

5. The method of claim 2, repeating steps (B) and (C) until a reference record is determined to exactly match the received data record, or until a predetermined maximum number of approximate matches reference records have been determined to approximately match the received data record.

6. The method of claim 2, wherein if a reference record is found that exactly matches the received data record, additionally comprising:

(D) performing automatic processing on the exact matching reference record and the received data record.

7. The method of claim 2, wherein if no reference records are determined to exactly match the received data record, and no reference records are determined to approximately match the received data record, storing the received data record as a persistent query.

8. The method of claim 1, wherein the system is an AP processing system, the received data record is an invoice, and the stored reference records are purchase orders.

9. The method of claim 1, additionally comprising:

(D) ranking the records that were determined to approximately match the received data record by the calculated distance.

10. The method of claim 9, wherein if the top-ranked approximately matching reference record is within a predefined automatic processing threshold, correcting the top-ranked approximately matching reference record to match the received data record and automatically processing the corrected reference record and the received data record.

11. The method of claim 9, wherein if only one reference record is determined to approximately match the received data record, and if the top-ranked approximately matching reference record is within a predefined automatic processing threshold, correcting the top-ranked approximately matching reference record to match the received data record and automatically processing the corrected reference record and the received data record.

12. The method of claim 1, wherein the matching system is bond trading system, and the received data record is a bond request, and the stored reference records are bonds available for purchase.

13. A computer-readable medium containing instructions that cause a computer to process a data record in a matching system, said medium comprising:

a storage section for storing a plurality of data records;

a first code section containing instructions that cause the computer to receive the data record;

a second code section containing instructions that cause the computer to calculate a distance between the received data record and a reference record; and

a third code section containing instructions that cause the computer to determine that the reference record approximately matches the received data record if the calculated distance is less than a pre-determined threshold.

14. The computer-readable medium of claim 13, wherein said third code section additionally contains instructions that cause the computer to determine that the reference record exactly matches the received data record if the calculated distance is zero.

15. The computer-readable medium of claim 14, additionally comprising a fourth code that contains instructions to repeat the instructions in the second code section and the third code section until a reference record is determined to exactly match the received data record.

16. The computer-readable medium of claim 14, additionally comprising a fourth code section that contains instructions to repeat the instructions in the second code section and the third code section until a reference record is determined to exactly match the received data record, or until a predetermined maximum number of distances have been calculated.

17. The computer-readable medium of claim 14, additionally comprising a fourth code section that contains instruction to repeat the instructions in the second code section and the third code section until a reference record is determined to exactly match the received data record, or until a predetermined maximum number of approximate matches reference records have been determined to approximately match the received data record.

18. The computer-readable medium of claim 14, wherein if a reference record is found that exactly matches the received data record, additionally comprising fourth code section containing instructions that cause a computer to perform automatic processing on the exact matching reference record and the received data record.

19. The computer-readable medium of claim 14, wherein if no reference records are determined to exactly match the received data record, and no reference records are determined to approximately match the received data record, a fourth computer code section containing instructions that cause the computer to store the received data record as a persistent query in the storage section.

20. The computer-readable medium of claim 13, wherein the system is an AP processing system, the received data record is an invoice, and the stored reference records are purchase orders.

21. The computer-readable medium of claim 13, additionally comprising a fourth code section containing instructions that cause the computer to rank the records that were determined to approximately match the received data record by the calculated distance.

22. The computer-readable medium of claim 21, wherein if the top-ranked approximately matching reference record is within a predefined automatic processing threshold, additionally comprising a fifth code section containing instructions that cause the computer to correct the top-ranked approximately matching reference record to match the received data record and automatically process the corrected reference record and the received data record.

23. The computer-readable medium of claim 21, wherein if the top-ranked approximately matching reference record is within a predefined automatic processing threshold, additionally comprising a fifth code section containing instructions that cause the computer to correct the top-ranked approximately matching reference record to match the received data record and automatically process the corrected reference record and the received data record.

24. The computer-readable medium of claim 23, wherein the matching system is bond trading system, and the received data record is a bond request, and the stored reference records are bonds available for purchase.

25. A method of determining a record distance between a reference record and an incoming record, wherein the reference record has at least one field and the incoming record has at least one field, comprising:

(A) determining correspondence between at least one field in the reference record to at least one field in the incoming data record;

(B) for each corresponding pair of fields determined in step (a), calculating a field distance; and

(C) calculating said record distance by aggregating said field distances.

26. The method of claim 25, wherein said reference record is a persistent query.

27. The method of claim 25, wherein step (C) additionally comprises:

(i) totaling each field distance determined in step (B).

28. The method of claim 25, wherein step (B) additionally comprises:

(i) for each corresponding pair of fields determined in step (A), calculating a field distance and weighing the calculated field distance by a field-specific predetermined weight to calculate a weighted field distance.

29. The method of claim 28, wherein step (C) additionally comprises:

(i) totaling each weighted field distance determined in step (B)(i).

30. The method of claim 29, wherein step (C)(i) additionally comprises totaling each weighted field distance determined in step (B) using a weighted sum of squares function.

31. The method of claim 25, wherein said field distance is calculated using a field distance function.

32. The method of claim 31, wherein a field distance function is defined for each corresponding pair of fields.

33. The method of claim 32, wherein at least one field distance function is predefined.

34. The method of claim 32, wherein at least one field distance function is user-defined.

35. The method of claim 25, wherein if a calculated record distance is zero, the reference record and the incoming record match.

36. A computer-readable medium for determining a record distance between a reference record and an incoming record, wherein the reference record has at least one field and the incoming record has at least one field, said medium comprising:

a first code section containing instructions that cause the computer to determine correspondence between at least one field in the reference record to at least one field in the incoming data record;

a second code section containing instructions that cause the computer to calculate a field distance for each corresponding pair of fields; and

a third code section containing instructions that cause the computer to calculate said record distance by aggregating said field distances.

37. The computer-readable medium of claim 36, wherein said reference record is a persistent query.

38. The computer-readable medium of claim 36, wherein said third code section additionally contains instructions that cause the computer to total each calculated field distance.

39. The computer-readable medium of claim 36, wherein said second code section additionally contains instructions that cause the computer to calculated weighted field distance by calculating a field distance for each corresponding pair of fields, and weighing the calculated field distance by a field-specific predetermined weight.

40. The computer-readable medium of claim 39, wherein said third code section additionally comprises instructions that cause the computer to total each weighted field distance.

41. The computer-readable medium of claim 40, wherein said third code section additionally comprises instructions that cause the computer to total each weighted field distance determined using a weighted sum of squares function.

42. The computer-readable medium of claim 36, wherein said second code section contains instructions that cause the computer to calculate the field distance using a field distance function.

43. The computer-readable medium of claim 42, wherein the field distance function used by the second code section is separately defined for each corresponding pair of fields.