Product recommendations based on characteristics from end user-generated text

Patent metadata

Field	Value
Publication number	US-12182849-B2
Application number	US-202418420362-A
Country	US
Kind code	B2
Filing date	Jan 23, 2024
Priority date	Jan 31, 2022
Publication date	Dec 31, 2024
Grant date	Dec 31, 2024

Abstract

A method for generating a related product interface portion in an electronic user interface based on user-generated text includes identifying, for each product of a plurality of products, one or more respective characteristics from end user-generated text associated with the product, whereby a plurality of characteristics are identified for the plurality of products. The method further includes determining a plurality of characteristic groups, each group comprising two or more of the plurality of characteristics, wherein the characteristics within a group are similar to each other; receiving, from a user through the electronic user interface, an input related to a product category; and in response to receiving the user input, generating and presenting the related product interface portion that includes a respective product from each of the two or more of the characteristic groups related to the category.

First claim

What is claimed is:

1. A method comprising:
(i) applying, to respective end user-generated text associated with each product of a plurality of products, a machine learning model trained to output portions of the end user-generated text that include a plurality of characteristics associated with the plurality of products;
(ii) generating a plurality of embeddings vectors, each of the plurality of embeddings vectors representative of a respective one of the plurality of characteristics;
(iii) identifying an initial vector of the plurality of embeddings vectors as having a shortest average distance to the plurality of embeddings vectors;
(iv) defining an initial cluster of embeddings vectors as the plurality of embeddings vectors within a threshold distance of the initial vector, the initial cluster being a first cluster in one or more clusters;
(v) identifying a further vector of the plurality of embeddings vectors having a shortest average distance to the plurality of embeddings vectors not in a cluster of the one or more clusters;
(vi) defining a further cluster of embeddings vectors as the plurality of embeddings vectors within a threshold distance of the further vector and adding the further cluster to the one or more clusters;
(vii) repeating (v) and (vi) until each of the plurality of embeddings vectors is within a cluster of the one or more clusters;
(viii) receiving an input related to a product category; and
(ix) generating and presenting a graphical user interface that includes a respective product from the defined cluster associated with the product category.

2. The method of claim 1 , wherein each of the plurality of characteristics comprises one or more plaintext words from the end user-generated text.

3. The method of claim 1 , further comprising: determining a respective representative word for each cluster of the one or more clusters; determining a pairwise similarity of each characteristic to the representative word for each cluster; and adding products to the plurality of clusters according to matches in the pairwise similarity.

4. The method of claim 3 , wherein the representative word of the first cluster corresponds to the first vector.

5. The method of claim 1 , further comprising: selecting, for each cluster of the one or more clusters, a respective representative product according to one or more of respective popularities of the products within the group or a confidence score associated with the determination of the characteristic from the end user-generated text associated with the product.

6. The method of claim 1 , wherein the plurality of products are within a same product category.

7. The method of claim 1 , wherein the machine learning model is a natural language processing model.

8. A system for generating a related product interface portion in an electronic user interface based on user text, the system comprising:
a non-transitory, computer-readable medium storing instructions; and
a processor configured to execute the instructions to:
(i) apply, to respective end user-generated text associated with each product of a plurality of products, a machine learning model trained to output portions of the end user-generated text that include a plurality of characteristics associated with the plurality of products;
(ii) generate a plurality of embeddings vectors, each of the plurality of embeddings vectors representative of a respective one of the plurality of characteristics;
(iii) identify an initial vector of the plurality of embeddings vectors as having a shortest average distance to the plurality of embeddings vectors;
(iv) define an initial cluster of embeddings vectors as the plurality of embeddings vectors within a threshold distance of the initial vector, the initial cluster being a first cluster in one or more clusters;
(v) identify a second vector of the plurality of embeddings vectors having a shortest average distance to the plurality of embeddings vectors not in the first cluster;
(vi) define a further cluster of embeddings vectors as the plurality of embeddings vectors within a threshold distance of the further vector and add the further cluster to the one or more clusters;
(vii) repeat steps (v) and (vi) until each of the plurality of embeddings vectors is within a cluster of the one or more clusters;
(viii) receive an input related to a product category; and
(ix) generate and present a graphical user interface that includes a respective product from the defined cluster associated with the product category.

9. The system of claim 8 , wherein each of the plurality of characteristics comprises one or more plaintext words from the end user-generated text.

10. The system of claim 8 , wherein the instructions further cause the processor to: determine a respective representative word for each cluster of the one or more clusters; determine a pairwise similarity of each characteristic to the representative word for each cluster; and add products to the plurality of clusters according to matches in the pairwise similarity.

11. The system of claim 10 , wherein the representative word of the first cluster corresponds to the first vector.

12. The system of claim 8 , wherein the instructions further cause the processor to: select, for each cluster of the one or more clusters, a respective representative product according to one or more of respective popularities of the products within the group or a confidence score associated with the determination of the characteristic from the end user-generated text associated with the product.

13. The system of claim 8 , wherein the plurality of products are within a same product category.

14. The system of claim 8 , wherein the machine learning model is a natural language processing model.

15. A non-transitory computer-readable storage medium storing program instructions that, when executed by a processor, cause a computer system to perform operations comprising:
(i) applying, to respective end user-generated text associated with each product of a plurality of products, a machine learning model trained to output portions of the end user-generated text that include a plurality of characteristics associated with the plurality of products;
(ii) generating a plurality of embeddings vectors, each of the plurality of embeddings vectors representative of a respective one of the plurality of characteristics;
(iii) identifying an initial vector of the plurality of embeddings vectors as having a shortest average distance to the plurality of embeddings vectors;
(iv) defining an initial cluster of embeddings vectors as the plurality of embeddings vectors within a threshold distance of the initial vector, the initial cluster being a first cluster in one or more clusters;
(v) identifying a further vector of the plurality of embeddings vectors having a shortest average distance to the plurality of embeddings vectors not in a cluster of the one or more clusters;
(vi) defining a further cluster of embeddings vectors as the plurality of embeddings vectors within a threshold distance of the further vector and adding the further cluster to the one or more clusters;
(vii) repeating steps (v) and (vi) until each of the plurality of embeddings vectors is within a cluster of the one or more clusters;
(viii) receiving an input related to a product category; and
(ix) generating and presenting a graphical user interface that includes a respective product from the defined cluster associated with the product category.

16. The non-transitory computer-readable storage medium of claim 15 , wherein each of the plurality of characteristi

Assignees

• Home Depot Product Authority Llc

Inventors

• Xu Tianlong
• Tian Haozheng
• Yan Nian
• Nair Harish

Classifications

• G06Q30/0282

  Rating or review of business operators or products · CPC title

• G06F40/40

  Processing or translation of natural language (natural language analysis G06F40/20; semantic analysis G06F40/30) · CPC title

• G06Q30/0627 Primary

  by specifying product or service characteristics, e.g. product dimensions · CPC title

• G06N3/08

  Learning methods · CPC title

• G06N3/0455

  Auto-encoder networks; Encoder-decoder networks · CPC title