Method and apparatus for correlating products based upon logical relations by product attributes and optionally with customer data purchases

• US 9,047,610 B1
• Filed: 08/04/2008
• Issued: 06/02/2015
• Est. Priority Date: 07/27/2000
• Status: Active Grant

• Current Assignee: AudienceScience, Inc.
• Original Assignee: AudienceScience, Inc.
• Inventors: Nayfeh, Basem, Chen, Cheng Che, Haddara, Yaser M., Kozyrczak, Maciek
• Primary Examiner(s): DURAN, ARTHUR D
• Application Number:
  US12/185,761
• Time in Patent Office: 2,493 Days
• US Class Current: 1/1
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

A method for generating a list of recommendations to be supplied by a vendor to a prospective customer by the use of a computer comprising the steps of inputting into a searchable data set a plurality of products and attributes of each product. A product is. The attributes of the selected product are ascertained. All products are ascertained that match at least one of the attributes of the selected product to provide looked-up products. From the looked-up products then are selected all products that are the closest match to the selected product to provide best matches. An output list of best matches of correlated products from the looked-up products is generated.

Claims

1. A method, performed by a computing system having a processor, for generating a list of at least one correlated product to be supplied by a vendor to a prospective customer by the use of a computer, the method comprising:

organizing, with the processor, into a searchable data set a plurality of products and attributes of each product;

selecting, with the processor, a product having a plurality of attributes;

ascertaining, with the processor, the plurality of attributes of the selected product;

for each of at least two of the plurality of attributes,

ascertaining, with the processor, all products that match the attribute of the selected product to provide looked-up products;

selecting, with the processor, from the looked-up products all products that are the closest match to the selected product utilizing predetermined attribute-related criteria to provide best matches;

generating, with the processor, an output list of at least one of the best matches from the looked-up products; and

outputting the generated output list to the prospective customer,

wherein the best matches are ascertained by utilizing predetermined attribute-related criteria and wherein one attribute-related criterion is all products that match on the one attribute that has the least number of products in it.

2. The method of claim 1 further comprising:

inserting purchase data of the products into data structures;

ascertaining all products that were purchased by a customer who also purchased the selected product to provide additional selected products;

ranking the additional selected products in accordance with a correlation index;

selecting from the additional selected products all products that match the selected product on at least one attribute; and

preparing an output list of at least one of the best matches.

3. The method of claim 2 comprising:

ranking the best matches for a selected product in accordance with a relative benefit to the vendor of selling each of the best matches for the selected product; and

preparing a list of recommendations for the selected product based upon the relative benefit ranking.

4. The method of claim 3 further comprising preparing the list with the top “

n”

products from the ranked best matches.

5. The method of claim 1 further comprising repeating the steps of obtaining an output list for each of a plurality of selected products.

6. The method of claim 1 further comprising preparing a look-up table for the attributes by product.

7. The method of claim 1 wherein one predetermined attribute-related criterion is all products that have the largest number of attributes in common with the selected product.

8. The method of claim 2 wherein the correlation index is determined by the number of customers who purchased both the selected product and the additional selected products.

9. The method of claim 2 wherein an estimator of the likelihood of purchase is calculated from the linear correlation between purchases of the selected product and the additional selected products.

10. The method of claim 9 wherein the correlation index is determined by a tradeoff of the mean and variance of the estimator of the likelihood that a customer who purchases the selected product would also purchase the additional selected products.

11. The method of claim 6 further comprising preparing a look-up table for the products by attribute.

12. The method of claim 11 wherein ascertaining the plurality of attributes of the selected product includes applying the look-up table for the attributes by product to the identity of the selected product;

and

wherein ascertaining all of the products that match at least one of the attributes of the selected product to provide looked-up products involves applying the look-up table for the products by attributes to the ascertained attributes of the selected product.

13. The method of claim 1 wherein a distinguished pair of the ascertained attributes of the selected product has a hierarchical relationship such that one attribute of the distinguished pair is a broadly based attribute, while the other attribute of the distinguished pair is a specific attribute;

and

wherein all products that match the broadly based attribute of the distinguished pair of attributes are ascertained before all products that match the specific attribute of the distinguished pair of attributes is ascertained.

14. The method of claim 1 wherein the ascertaining of all products that match an attribute of the selected product to provide looked-up products is performed for each of the plurality of attributes.

15. The method of claim 1, further comprising:

ranking the best matches for a selected product in accordance with a relative benefit to the vendor of selling each of the best matches for the selected product; and

preparing a list of recommendations for the selected product based upon the relative benefit ranking.

16. The method of claim 1, further comprising:

preparing a look-up table for the attributes by product; and

preparing a look-up table for the products by attribute,

and wherein ascertaining the attributes of the selected products includes applying the look-up table for the attributes by product to the identity of the selected product, and wherein ascertaining all of the products that match at least one of the attributes of the selected product to provide looked-up products involves applying the look-up table for the products by attributes to the ascertained attributes of the selected product.

17. The method of claim 1, wherein a distinguished pair of the ascertained attributes of the selected product has a hierarchical relationship such that one attribute of the distinguished pair is a broadly based attribute, while the other attribute of the distinguished pair is a specific attribute;

and

wherein all products that match the broadly based attribute of the distinguished pair of attributes are ascertained before all products that match the specific attribute of the distinguished pair of attributes are ascertained.

18. The method of claim 1, further comprising:

ascertaining of all products that match an attribute of the selected product for each of the plurality of attributes.

19. A computer-readable device, that is not a signal, whose contents cause a computing system to generate a list of at least one correlated product to be supplied by a vendor to a prospective customer by:

organizing into a searchable data set a plurality of products and attributes of each product;

selecting a product having a plurality of attributes;

ascertaining the attributes of the selected product;

for each of at least two of the plurality of attributes,

ascertaining all products that match the attribute of the selected product to provide looked-up products;

selecting from the looked-up products all products that are the closest match to the selected product utilizing predetermined attribute-related criteria to provide best matches;

generating an output list of at least one of the best matches from the looked-up products; and

output the generated output list to the prospective customer,

wherein the best matches are ascertained by utilizing predetermined attribute-related criteria and wherein one attribute-related criterion is all products that match on the one attribute that has the least number of products in it.

20. The computer-readable device, that is not a signal, of claim 19 further comprising contents that cause the computing system to:

rank the best matches for a selected product in accordance with a relative benefit to the vendor of selling each of the best matches for the selected product; and

prepare a list of recommendations for the selected product based upon the relative benefit ranking.

21. The computer-readable device, that is not a signal, of claim 19 further comprising contents that cause the computing system to:

prepare a look-up table for the attributes by product; and

prepare a look-up table for the products by attribute,

and wherein ascertaining the attributes of the selected products includes applying the look-up table for the attributes by product to the identity of the selected product, and wherein ascertaining all of the products that match at least one of the attributes of the selected product to provide looked-up products involves applying the look-up table for the products by attributes to the ascertained attributes of the selected product.

22. The computer-readable device, that is not a signal, of claim 19 wherein a distinguished pair of the ascertained attributes of the selected product has a hierarchical relationship such that one attribute of the distinguished pair is a broadly based attribute, while the other attribute of the distinguished pair is a specific attribute;

and

wherein all products that match the broadly based attribute of the distinguished pair of attributes are ascertained before all products that match the specific attribute of the distinguished pair of attributes are ascertained.

23. The computer-readable device, that is not a signal, of claim 19 further comprising contents that cause the computing system to perform the ascertaining of all products that match an attribute of the selected product for each of the plurality of attributes.

24. The computer-readable device, that is not a signal, of claim 19 whose contents are further configured to cause the computing system to generate the list of at least one correlated product to be supplied by the vendor to the prospective customer by:

inserting purchase data of the products into data structures;

ascertaining all products that were purchased by a customer who also purchased the selected product to provide additional selected products;

ranking the additional selected products in accordance with a correlation index;

selecting from the additional selected products all products that match the selected product on at least one attribute; and

preparing an output list of at least one of the best matches.

25. The computer-readable device, that is not a signal, of claim 24 whose contents are further configured to cause the computing system to generate the list of at least one correlated product to be supplied by the vendor to the prospective customer by:

ranking the best matches for a selected product in accordance with a relative benefit to the vendor of selling each of the best matches for the selected product; and

preparing a list of recommendations for the selected product based upon the relative benefit ranking.

26. The computer-readable device, that is not a signal, of claim 19 whose contents are further configured to cause the computing system to generate the list of at least one correlated product to be supplied by the vendor to the prospective customer by:

preparing a look-up table for the attributes by product.

27. The computer-readable device, that is not a signal, of claim 19, wherein one predetermined attribute-related criterion is all products that have the largest number of attributes in common with the selected product.

28. The computer-readable device, that is not a signal, of claim 24, wherein the correlation index is determined by the number of customers who purchased both the selected product and the additional selected products.

29. The computer-readable device, that is not a signal, of claim 26 whose contents are further configured to cause the computing system to generate the list of at least one correlated product to be supplied by the vendor to the prospective customer by:

preparing a look-up table for the products by attribute.

30. The computer-readable device, that is not a signal, of claim 29 wherein ascertaining the plurality of attributes of the selected product includes applying the look-up table for the attributes by product to the identity of the selected product;

and

wherein ascertaining all of the products that match at least one of the attributes of the selected product to provide looked-up products involves applying the look-up table for the products by attributes to the ascertained attributes of the selected product.

31. A computing system for generating a list of at least one correlated product to be supplied by a vendor to a prospective customer by the use of a computer, the computing system comprising:

a memory; and

one or more processors configured to;

organize into a searchable data set a plurality of products and attributes of each product;

select a product having a plurality of attributes;

ascertain the plurality of attributes of the selected product;

for each of at least two of the plurality of attributes,

ascertain all products that match the attribute of the selected product to provide looked-up products;

select from the looked-up products all products that are the closest match to the selected product utilizing predetermined attribute-related criteria to provide best matches;

generate an output list of at least one of the best matches from the looked-up products, and

output the generated output list to the prospective customer,

wherein the best matches are ascertained by utilizing predetermined attribute-related criteria and wherein one attribute-related criterion is all products that match on the one attribute that has the least number of products in it.

32. The computing system of claim 31, wherein the one or more processors are further configured to:

rank the best matches for a selected product in accordance with a relative benefit to the vendor of selling each of the best matches for the selected product; and

prepare a list of recommendations for the selected product based upon the relative benefit ranking.

33. The computing system of claim 31, wherein the one or more processors are further configured to:

prepare a look-up table for the attributes by product; and

prepare a look-up table for the products by attribute,

and wherein the one or more processors configured to ascertain the attributes of the selected products apply the look-up table for the attributes by product to the identity of the selected product, and wherein the one or more processors configured to ascertain all of the products that match at least one of the attributes of the selected product to provide looked-up products apply the look-up table for the products by attributes to the ascertained attributes of the selected product.

34. The computing system of claim 31, wherein the one or more processors are further configured to:

ascertain all products that match an attribute of the selected product for each of the plurality of attributes.

35. The computing system of claim 31, wherein the one or more processors are further configured to:

insert purchase data of the products into data structures;

ascertain all products that were purchased by a customer who also purchased the selected product to provide additional selected products;

rank the additional selected products in accordance with a correlation index;

select from the additional selected products all products that match the selected product on at least one attribute; and

prepare an output list of at least one of the best matches.