Automatic product categorization

• US 7,870,039 B1
• Filed: 08/17/2004
• Issued: 01/11/2011
• Est. Priority Date: 02/27/2004
• Status: Active Grant

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Pavlov, Dmitry, Suresh, Bipin, Dom, Byron Edward, Balasubramanyan, Ramnath, Goyal, Abhishek
• Primary Examiner(s): DUNHAM, JASON B
• Application Number:
  US10/920,588
• Time in Patent Office: 2,338 Days
• Field of Search: 705/26, 705/27
• US Class Current: 705/26.41
• CPC Class Codes: G06F 16/355 Class or cluster creation or modification
• CPC Class Codes: G06Q 30/0613 Third-party assisted
• CPC Class Codes: G06Q 30/0623 Item investigation

Abstract

Techniques are provided for automatic product categorization. In one aspect, the categorization is based on text and one or more other values associated with a product offering. In another aspect, a first categorization of a product offering is performed and, if the product category chosen is in a set of co-refinable product categories, then a second (or third, fourth and so on) categorization is performed among the set of co-refinable product categories. In a third aspect, products are categorized based on cost. In another aspect, after products are categorized, the products are flagged for further categorization processing if the cost for categorizing the product is beyond a predefined threshold.

Claims

1. A method for automatically categorizing a product offering comprising the machine-implemented steps of:

obtaining data associated with the product offering of a product;

wherein the data includes a plurality of fields, wherein the plurality of fields include text of the product offering and one or more values associated with the product offering;

automatically determining a particular product category, of a plurality of product categories, to which to first assign the product offering, responsive to a determination of an existence of a particular field within the plurality of fields without regard to the value of the particular field; and

responsive to a determination that the particular product category belongs to a set of co-refinable product categories, performing a second product category assignment,

wherein the second product category assignment is performed by determining which product category, of the set of co-refinable product categories, to assign to the product offering;

wherein each category in the set of co-refinable product categories is also one of the plurality of product categories and wherein one or more categories of the plurality of product categories is not in the set of co-refinable product categories; and

wherein the method is performed by one or more computing devices comprising one or more processors.

2. The method of claim 1, wherein the one or more values comprise one or more of a merchant identifier for the product offering, brand information for the product offering, and a price of the product offering.

3. The method of claim 1, wherein the set of co-refinable product categories is determined based on discrepancies between results of two or more different assignment methods.

4. The method of claim 1, wherein the text of the product offering comprises one or more of a title associated with the product offering and a human-readable description associated with the product offering.

5. The method of claim 1, wherein the set of co-refinable product categories is determined manually.

6. The method of claim 1, wherein one or both of the first and second assignments are determined, at least in part, by a Bayesian method.

7. The method of claim 1, wherein one or both of the first and second assignments includes assigning a plurality of product offerings.

8. The method of claim 1,

wherein the determination for the second assignment is based on a different set of fields within the plurality of fields than the determination for the first assignment.

9. The method of claim 1, wherein the determination for the second assignment is based on a different assignment method than the determination for the first assignment.

10. The method of claim 1, wherein the determination for the second assignment is more computationally expensive than the determination for the first assignment.

11. The method of claim 1, wherein one or both of the first and second assignments are based, at least in part, on one or more of assignment-error costs, results of a confusion matrix, likelihood that a given product offering would be mistakenly categorized into each other product category in the plurality of categories, and an existence of a particular field within the plurality of fields without regard to the value of that particular field.

12. The method of claim 1, wherein the co-refinable product category sets are based on the likelihood of incorrectly assigning between the members of the co-refinable product category set.

13. The method of claim 1, wherein one or more values are distinct and separate from the text of the product offering.

14. The method of claim 1, wherein the determination for the second assignment is based, at least in part, only on categories which are members of the co-refinable set.

15. A non-transitory computer-readable storage medium storing instructions, the instructions including instructions for automatically categorizing a product offering which, when executed by one or more processors, cause the one or more processors to perform the steps of:

obtaining data associated with the product offering of a product;

wherein the data includes a plurality of fields, wherein the plurality of fields include text of the product offering and one or more values associated with the product offering;

automatically determining a particular product category, of a plurality of product categories, to which to first assign the product offering, responsive to a determination of an existence of a particular field within the plurality of fields without regard to the value of the particular field; and

responsive to a determination that the particular product category belongs to a set of co-refinable product categories, performing a second product category assignment,

wherein the second product category assignment is performed by determining which product category, of the set of co-refinable product categories, to assign to the product offering;

wherein each category in the set of co-refinable product categories is also one of the plurality of product categories and wherein one or more categories of the plurality of product categories is not in the set of co-refinable product categories.

16. The non-transitory computer-readable storage medium of claim 15, wherein the one or more values comprise one or more of a merchant identifier for the product offering, brand information for the product offering, and a price of the product offering.

17. The non-transitory computer-readable storage medium of claim 15, wherein the set of co-refinable product categories is determined based on discrepancies between results of two or more different assignment methods.

18. The non-transitory computer-readable storage medium of claim 15, wherein the text of the product offering comprises one or more of a title associated with the product offering and a human-readable description associated with the product offering.

19. The non-transitory computer-readable storage medium of claim 15, wherein the set of co-refinable product categories is determined manually.

20. The non-transitory computer-readable storage medium of claim 15, wherein one or both of the first and second assignments are determined, at least in part, by a Bayesian method.

21. The non-transitory computer-readable storage medium of claim 15, wherein one or both of the first and second assignments includes assigning a plurality of product offerings.

22. The non-transitory computer-readable storage medium of claim 15,

wherein the determination for the second assignment is based on a different set of fields within the plurality of fields than the determination for the first assignment.

23. The non-transitory computer-readable storage medium of claim 15, wherein the determination for the second assignment is based on a different assignment method than the determination for the first assignment.

24. The non-transitory computer-readable storage medium of claim 15, wherein the determination for the second assignment is more computationally expensive than the determination for the first assignment.

25. The non-transitory computer-readable storage medium of claim 15, wherein one or both of the first and second assignments are based, at least in part, on one or more of assignment-error costs, results of a confusion matrix, likelihood that a given product offering would be mistakenly categorized into each other product category in the plurality of categories, and an existence of a particular field within the plurality of fields without regard to the value of that particular field.

26. The non-transitory computer-readable storage medium of claim 15, wherein the co-refinable product category sets are based on the likelihood of incorrectly assigning between the members of the co-refinable product category set.

27. The non-transitory computer-readable storage medium of claim 15, wherein one or more values are distinct and separate from the text of the product offering.

28. The non-transitory computer-readable storage medium of claim 15, wherein the determination for the second assignment is based, at least in part, only on categories which are members of the co-refinable set.

29. An apparatus for automatically categorizing a product offering, comprising:

one or more processors; and

a non-transitory computer-readable medium storing instructions which, when processed by the one or more processors, causes the one or more processors to perform the steps of;

obtaining data associated with the product offering of a product;

wherein the data includes a plurality of fields, wherein the plurality of fields include text of the product offering and one or more values associated with the product offering;

automatically determining a particular product category, of a plurality of product categories, to which to first assign the product offering; and

determining that the particular product category belongs to a set of co-refinable product categories, and performing a second product category assignment,

wherein the second product category assignment is performed by determining which product category, of the set of co-refinable product categories, to assign to the product offering;

wherein the set of co-refinable product categories is a strict subset of the plurality of product categories.

30. The apparatus of claim 29, wherein the one or more values comprise one or more of a merchant identifier for the product offering, brand information for the product offering, and a price of the product offering.

31. The apparatus of claim 29, wherein the set of co-refinable product categories is determined based on discrepancies between results of two or more different assignment methods.

32. The apparatus of claim 29, wherein the text of the product offering comprises one or more of a title associated with the product offering and a human-readable description associated with the product offering.

33. The apparatus of claim 29, wherein the set of co-refinable product categories is determined manually.

34. The apparatus of claim 29, wherein one or both of the first and second assignments are determined, at least in part, by a Bayesian method.

35. The apparatus of claim 29, wherein one or both of the first and second assignments includes assigning a plurality of product offerings.

36. The apparatus of claim 29, wherein the determination for the second assignment is based on a different set of fields within the plurality of fields than the determination for the first assignment.

37. The apparatus of claim 29, wherein the determination for the second assignment is based on a different assignment method than the determination for the first assignment.

38. The apparatus of claim 29, wherein the determination for the second assignment is more computationally expensive than the determination for the first assignment.

39. The apparatus of claim 29, wherein one or both of the first and second assignments are based, at least in part, on one or more of assignment-error costs, results of a confusion matrix, likelihood that a given product offering would be mistakenly categorized into each other product category in the plurality of categories, and an existence of a particular field within the plurality of fields without regard to the value of that particular field.

40. The apparatus of claim 29, wherein the co-refinable product category sets are based on the likelihood of incorrectly assigning between the members of the co-refinable product category set.

41. The apparatus of claim 29, wherein one or more values are distinct and separate from the text of the product offering.

42. The apparatus of claim 29, wherein the determination for the second assignment is based, at least in part, only on categories which are members of the co-refinable set.