Methods for product collection recommendations based on transaction data

Patent metadata

Field	Value
Publication number	US-11861675-B2
Application number	US-202016785104-A
Country	US
Kind code	B2
Filing date	Feb 7, 2020
Priority date	Apr 22, 2019
Publication date	Jan 2, 2024
Grant date	Jan 2, 2024

Abstract

A method includes determining a first taxonomy of an anchor product. The first taxonomy includes a plurality of levels for classifying products organized from a highest taxonomy level to a lowest taxonomy level. The method further includes determining a second taxonomy closest to the first taxonomy. The second taxonomy is associated with a group of products, the first taxonomy and the second taxonomy have at least a common highest taxonomy level, and the determination is made at least in part based on co-purchase data indicating that the anchor product and at least one product in the group of products are purchased together more often than products associated with other taxonomies are purchased with the anchor product. The method further includes determining a most similar product to the anchor product from the group of products of the second taxonomy and associating the anchor product and the most similar product with one another in a product collection.

First claim

What is claimed is:

1. A computer-implemented method, comprising:
determining, by one or more processors of one or more computing devices associated with a retailer website, a first taxonomy of an anchor product, wherein:
the first taxonomy comprises a plurality of levels for classifying products organized from a highest taxonomy level to a lowest taxonomy level; and
a greater number of products are classified under the highest taxonomy level than under the lowest taxonomy level;

determining, by the one or more processors, a second taxonomy closest to the first taxonomy, wherein:
the second taxonomy is associated with a group of products,
the first taxonomy and the second taxonomy have at least a common highest taxonomy level,
each of the first taxonomy of the anchor product and the second taxonomy includes at least three levels,
the determination that the second taxonomy is closest to the first taxonomy is made at least in part based on co-purchase data indicating that the anchor product and at least one product in the group of products are purchased together more often than products associated with other taxonomies are purchased with the anchor product,
the determination that the second taxonomy is closest to the first taxonomy further comprises determining that at least the two highest levels of the first taxonomy and the second taxonomy are exactly the same, and
the first taxonomy and the second taxonomy have at least different lowest taxonomy levels;

determining, by the one or more processors, a most similar product to the anchor product from a group of products associated with the lowest level of the second taxonomy, such that the first taxonomy of the anchor product and the second taxonomy of the most similar product have at least the two highest levels that are exactly the same and the lowest levels that are different, wherein the determining the most similar product from the group of products further comprises:
training a neural network using manually-created sets of similar products;
inputting product information of the anchor product and product information of each product of the group of products into the trained neural network;
performing, using a cosine similarity layer of the trained neural network, a comparison between product information of the anchor product and product information of each product of the group of products, the comparison comprising applying a cosine similarity function to at least one vector of the anchor product and at least one vector of each product of the group of products; and
determining, based on the comparison, that the product information of the anchor product is closest to product information of the most similar product;

associating, by the one or more processors, the anchor product and the most similar product with one another in a product collection, and
sending, by the one or more processors to a user computing device, data structured such that the most similar product is displayed on a display of the user computing device along with the anchor product on a display of the user computing device as part of a single interface or webpage of the retailer website provided by the one or more computing devices associated with the retailer website.

2. The computer-implemented method of claim 1 , further comprising:
receiving a selection of the anchor product from a user electronic device;
determining that the most similar product is in the product collection with the anchor product; and
sending, to the user electronic device, information for presenting the most similar product along with the anchor product on a display of the user electronic device responsive to the selection of the anchor product.

3. The computer-implemented method of claim 1 , wherein the comparison comprises performing a cosine similarity comparison between text of the product information of the anchor product and text of the product information of each product of the group of products.

4. The computer-implemented method of claim 1 , wherein the determination that the second taxonomy is closest to the first taxonomy further comprises:
determining a plurality of taxonomies for which each of a plurality of levels of each of the plurality of taxonomies is the same as the plurality of levels of the first taxonomy except the lowest taxonomy level, wherein the second taxonomy is one of the plurality of taxonomies; and
determining that the group of products associated with the second taxonomy has been purchased together with the anchor product more frequently than any other group of products associated with the remaining plurality of taxonomies.

5. The computer-implemented method of claim 1 , wherein the determination that the second taxonomy is closest to the first taxonomy is made at least in part based on co-view data indicating that the anchor product and the at least one product in the group of products are viewed by users during a single session more often than groups of products associated with other taxonomies.

6. The computer-implemented method of claim 1 , further comprising determining a third product to associate with the anchor product in the product collection, wherein the third product is selected from a second group of products associated with a third taxonomy that is a next closest to the first taxonomy after the second taxonomy.

7. The computer-implemented method of claim 6 , further comprising:
receiving a selection of the anchor product from a user electronic device;
determining that the most similar product and the third product are both in the product collection with the anchor product; and
sending, to the user electronic device, information for presenting the most similar product and the third product along with the anchor product on a display of the user electronic device responsive to the selection of the anchor product.

8. The computer-implemented method of claim 1 , further comprising:
determining a plurality of taxonomies for which a predetermined number of a plurality of levels of each of the plurality of taxonomies is the same as the plurality of levels of the first taxonomy;
determining the second taxonomy from the plurality of taxonomies by:
determining, based on co-purchase data, a percentage for each of the plurality of taxonomies that any product in the first taxonomy has been co-purchased with any product of each of the respective plurality of taxonomies, and
determining that the second taxonomy has the highest percentage of products that have been co-purchased with products of the first taxonomy.

9. The computer-implemented method of claim 8 , further comprising:
receiving a selection of the anchor product from a user electronic device;
determining that the most similar product is in the product collection with the anchor product; and
sending, to the user electronic device, information for presenting the most similar product along with the anchor product on a display of the user electronic device responsive to the selection of the anchor product.

10. The computer-implemented method of claim 8 , wherein determining the most similar product from the group of products further comprises:
performing a comparison between product information of the anchor product and product information of each product of the group of products; and
determining, based on the comparison, that the product information of the anchor product is closest to product information of the most similar product.

11. The computer-implemented method of claim 10 , wherein the comparison comprises performing a cosine similarity function between text of the product information of the anchor product and text of the product information of each product of the

Assignees

• Home Depot Product Authority Llc

Inventors

• Al Jadda Khalifeh
• Qu Huiming
• Yan Nian
• Ahsan Unaiza
• HE WU San

Classifications

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06Q30/0631 Primary

  Recommending goods or services · CPC title

• G06F16/285

  Clustering or classification · CPC title

• G06F16/9535

  Search customisation based on user profiles and personalisation · CPC title