Systems and methods for garment size recommendation

Patent metadata

Field	Value
Publication number	US-11727466-B2
Application number	US-202217664621-A
Country	US
Kind code	B2
Filing date	May 23, 2022
Priority date	Apr 29, 2020
Publication date	Aug 15, 2023
Grant date	Aug 15, 2023

Abstract

Disclosed are methods, systems, and non-transitory computer-readable medium for generating recommendations regarding products. A method may include determining a set of content features including one or more product attributes; determining a set of latent features; receiving a query user identifier and a query product identifier; determining a feature vector associated with the query user identifier and the query product identifier based on the set of content features and the set of latent features; determining one or more model coefficients for a linear model; and utilizing the linear model to determine a fit score for the query user identifier and the query product identifier.

First claim

What is claimed is:

1. A computer-implemented method for generating at least one fit recommendation for at least one article, the method comprising:
sending a size adviser API call in response to a user accessing a product page displayed on a user interface, wherein the size adviser API call includes a user identifier associated with the user and at least one unique article identifier;
in response to the size adviser API call, obtaining feedback or purchase information associated with the user, the feedback or purchase information including a plurality of data points;
comparing the plurality of data points with a predetermined threshold, including aggregating the plurality of data points;
in response to the comparing, determining that the aggregated plurality of data points exceed the predetermined threshold;
in response to determining that the aggregated plurality of data points exceed the predetermined threshold, determining a fit score for each of the at least one unique article identifier;
analyzing the fit score for each of the plurality of data points to determine a highest fit score corresponding to the at least one unique article identifier;
comparing measurements for an article associated with the highest fit score to a range of user measurements to determine if the measurements for the article associated with the highest fit score fall within the range of user measurements, the range of user measurements based on a profile of the user; and
in response to determining that the highest fit score falls within the range of user measurements, recommending the article associated with the highest fit score to the user via the user interface.

2. The computer-implemented method of claim 1 , the method further comprising:
determining training data based on the feedback or purchase information.

3. The computer-implemented method of claim 2 , wherein a parameterized model is trained based on the training data.

4. The computer-implemented method of claim 3 , the method further comprising:
using the parameterized model to determine the fit score.

5. The computer-implemented method of claim 1 , wherein the at least one unique article identifier corresponds to the at least one article depicted in the product page.

6. The computer-implemented method of claim 1 , the plurality of data points corresponding to at least one instance where the user previously provided feedback regarding an article or purchased the article.

7. The computer-implemented method of claim 1 , the method further comprising:
in response to determining that the aggregated plurality of data points do not exceed the predetermined threshold, using a content-based approach to determine a size recommendation.

8. The computer-implemented method of claim 7 , wherein the content-based approach includes determining a recommended unique article identifier based on a user profile or at least one garment attribute.

9. The computer-implemented method of claim 1 , the method further comprising:
in response to determining that the measurements for the article associated with the highest fit score do not fall within the range of user measurements, using a content-based approach to determine a recommended unique article identifier.

10. A computer system for generating at least one fit recommendation for at least one article comprising:
a data storage device storing processor-readable instructions; and
a processor configured to execute the instructions to perform a method including:
sending a size adviser API call in response to a user accessing a product page displayed on a user interface, wherein the size adviser API call includes a user identifier associated with the user and at least one unique article identifier;
in response to the size adviser API call, obtaining feedback or purchase information associated with the user, the feedback or purchase information including a plurality of data points;
comparing the plurality of data points with a predetermined threshold, including aggregating the plurality of data points;
in response to the comparing, determining that the aggregated plurality of data points exceed the predetermined threshold;
in response to determining that the aggregated plurality of data points exceed the predetermined threshold, determining a fit score for each of the at least one unique article identifier;
analyzing the fit score for each of the at least one unique article identifier to determine a highest fit score corresponding to the at least one unique article identifier;
comparing measurements for an article associated with the highest fit score to a range of user measurements to determine if the measurements for the article associated with the highest fit score fall within the range of user measurements, the range of user measurements based on a profile of the user; and
in response to determining that the highest fit score falls within the range of user measurements, recommending the article associated with the highest fit score to the user via the user interface.

11. The computer system of claim 10 , the method further comprising:
determining training data based on the feedback or purchase information.

12. The computer system of claim 11 , wherein a parameterized model is trained based on the training data.

13. The computer system of claim 12 , the method further comprising:
using the parameterized model to determine the fit score.

14. The computer system of claim 10 , wherein the at least one unique article identifier corresponds to the at least one article depicted in the product page.

15. The computer system of claim 10 , the plurality of data points corresponding to at least one instance where the user previously provided feedback regarding an article or purchased the article.

16. The computer system of claim 10 , the method further comprising:
in response to determining that the aggregated plurality of data points do not exceed the predetermined threshold, using a content-based approach to determine a size recommendation.

17. The computer system of claim 16 , wherein the content-based approach includes determining a recommended unique article identifier based on a user profile or at least one garment attribute.

18. A non-transitory computer-readable medium containing instructions that, when executed by a processor, cause the processor to perform a method for generating at least one fit recommendation for at least one article, the method comprising:
sending a size adviser API call, in response to a user accessing a product page displayed on a user interface, wherein the size adviser API call includes a user identifier associated with the user and at least one unique article identifier;
in response to the size adviser API call, obtaining feedback or purchase information associated with the user, the feedback or purchase information including a plurality of data points;
comparing the plurality of data points with a predetermined threshold, including aggregating the plurality of data points;
in response to the comparing, determining that the aggregated plurality of data points exceed the predetermined threshold;
in response to determining that the aggregated plurality of data points exceed the predetermined threshold, determining a fit score for each of the at least one unique article identifier;
analyzing the fit score for each of the at least one unique article identifier to determine a highest fit score corresponding to the at least one unique article identifier;
comparing measurements for an article associated with the highest fit score to

Assignees

• Caastle Inc

Inventors

• Zhang Fujun
• Jiang Dongming

Classifications

• G06Q30/0631 Primary

  Recommending goods or services · CPC title

• G06F16/9535

  Search customisation based on user profiles and personalisation · CPC title

• G06N5/04

  Inference or reasoning models · CPC title

• G06N20/00

  Machine learning · CPC title

• G06Q10/087

  Inventory or stock management, e.g. order filling, procurement or balancing against orders · CPC title