System and method for automatic product matching

• US 10,664,534 B2
• Filed: 03/15/2013
• Issued: 05/26/2020
• Est. Priority Date: 11/14/2012
• Status: Active Grant

• Current Assignee: Home Depot Product Authority, LLC (The Home Depot, Inc.)
• Original Assignee: Home Depot Product Authority, LLC (The Home Depot, Inc.)
• Inventors: Surya, Siva Kalyana Pavan Kumar Mallapragada Naga, Barcelo, Carlos
• Primary Examiner(s): Skhoun, Hicham
• Application Number:
  US13/838,195
• Publication Number:
  US 20140136549A1
• Time in Patent Office: 2,629 Days
• Field of Search: 707749
• CPC Class Codes: G06F 16/9532 Query formulation
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation

Abstract

A source product and a competitor product are automatically matched by computing similarity scores between the products. An attribute of a source product is extracted from information received from a user. The extracted attribute is then used to generate a search term. A search for competitor product information is performed using the search term and a similarity score between the source product and the competitor product is calculated. In one implementation, the similarity score is calculated by generating a first value pair for the source product and the competitor product, and for the first value pair, (1) assigning a value to a degree of similarity between the source product and the competitor product with respect to a common feature, and (2) finding the product of the assigned value and a weight value assigned to the common feature.

Claims

1. A method for automatically matching products, the method comprising:

extracting at least one attribute of a source product from unstructured source product information received from a user, the extracting comprising identifying at least one tag in the unstructured source product information, wherein the at least one tag is associated with the at least one attribute of the source product and a priority of the at least one attribute within a product group to which the source product belongs;

generating a search term from the at least one tag, wherein the search term comprises the at least one attribute;

using the search term to search for competitor product information; and

computing a similarity score between the source product information and the competitor product information, wherein computing the similarity score comprises;

generating a first value pair for the source product and a first competitor product, and for the first value pair;

assigning a value to a degree of similarity between the source product and the first competitor product with respect to a first common feature; and

finding the product of the assigned value and a weight value assigned to the first common feature.

2. The method of claim 1, wherein computing the similarity score further comprises:

for the first value pair;

assigning a second value to a degree of similarity between the source product and the first competitor product with respect to a second common feature;

finding the product of the assigned second value and a weight value assigned to the second common feature; and

summing the product for the first common feature and the product for the second common feature.

3. The method of claim 1, wherein computing the similarity score further comprises:

generating a second value pair for the source product and a second competitor product, and for the second value pair;

assigning a second value pair value to a degree of similarity between the source product and the second competitor product with respect to a feature common to the source product and second competitor product; and

finding the product of the assigned second value pair value and a weight value assigned to the feature common to the source product and the second competitor product.

4. The method of claim 1, wherein the weight value assigned to the first common feature is computed according to a logistic regression model.

5. The method of claim 1, wherein generating the search term from the at least one tag comprises:

generating the search term according to the priority.

6. The method of claim 1, wherein generating a first value pair for the source product and a first competitor product comprises:

converting the first value pair into a vector, the vector representing similarity between the information relating to the source produce and the first competitor product.

7. The method of claim 1, further comprising:

filtering out at least one value pair that violates a must match constraint.

15. The method of claim 1, further comprising:

extracting at least a second attribute of the source product from the unstructured source product information received from the user, the extracting comprising identifying at least a second tag in the unstructured source product information, wherein the at least a second tag is associated with the at least a second attribute of the source product and a priority of the at least a second attribute within the product group to which the source product belongs;

wherein generating a search term comprises generating a search term from the at least one tag and the at least a second tag, wherein the search term comprises the at least a first attribute and the at least a second attribute.

16. The method of claim 15, wherein the search term is a first search term, wherein generating the search term further comprises:

generating a second search term comprising one of the at least one attribute or the at least a second attribute, but not the other of the at least one attribute or the at least a second attribute, based on the respective priorities associated with at least at least one tag and the at least a second tag.

17. The method of claim 1, wherein the priority associated with the at least one tag is defined according to a non-binary schema.

8. A system for automatically matching products, the system comprising:

a memory; and

a processor coupled to the memory, the processor configured to;

extract a plurality of attributes of a source product from source product information received from a user, the extracting comprising identifying a plurality of tags in the source product information, wherein each of the plurality of tags is associated with a respective one of the plurality of attributes of the source product and a respective priority of the respective one of the plurality of attributes within a product group to which the source product belongs;

use a first search query generated from the plurality of tags to search for competitor product information, wherein the first search query comprises a first subset of the plurality of attributes;

use a second search query generated from the plurality of tags to search for competitor product information, wherein the second search query comprises a second subset of the plurality of attributes that is different from the first subset, wherein the first search query comprises information that is more reliable than information in the second search query; and

compute a similarity score between the source product information and the competitor product information, wherein computing the similarity score comprises;

generate a first value pair for the source product and a first competitor product, and for the first value pair;

assign a value to a degree of similarity between the source product and the first competitor product with respect to a first common feature; and

find the product of the assigned value and a weight value assigned to the first common feature.

9. The system of claim 8, wherein the processor is further configured to:

for the first value pair;

assign a second value to a degree of similarity between the source product and the first competitor product with respect to a second common feature;

find the product of the assigned second value and a weight value assigned to the second common feature; and

sum the product for the first common feature and the product for the second common feature.

10. The system of claim 8, wherein the processor is further configured to:

generate a second value pair for the source product and a second competitor product, and for the second value pair;

assign a second value pair value to a degree of similarity between the source product and the second competitor product with respect to a feature common to the source product and second competitor product; and

find the product of the assigned second value pair value and a weight value assigned to the feature common to the source product and the second competitor product.

11. The system of claim 8, wherein the weight value assigned to the first common feature is computed according to a logistic regression model.

12. The system of claim 8, wherein the processor is further configured to:

generate the first search query according to the respective priorities associated with the plurality of tags.

13. The system of claim 8, wherein the processor is further configured to:

convert the first value pair into a vector, the vector representing similarity between the information relating to the source produce and the first competitor product.

14. The system of claim 8, wherein the processor is further configured to:

filter out at least one value pair that violates a must match constraint.

18. The system according to claim 8, the processor further configured to:

extract at least a second attribute of the source product from the unstructured source product information received from the user, the extracting comprising identifying at least a second tag in the unstructured source product information, wherein the at least a second tag is associated with the at least a second attribute of the source product and a priority of the at least a second attribute within the product group to which the source product belongs;

wherein generating a search term comprises generating a search term from the at least one tag and the at least a second tag.

19. The system of claim 18, wherein generating the search term comprises:

generating a first search term comprising the at least one attribute and the at least a second attribute; and

generating a second search term comprising one of the at least one attribute or the at least a second attribute, but not the other of the at least one attribute or the at least a second attribute, based on the respective priorities associated with at least at least one tag and the at least a second tag.

20. The system of claim 8, wherein the priority associated with the at least one tag is defined according to a non-binary schema.