System, method, and medium to select a product title

Patent metadata

Field	Value
Publication number	US-11580589-B2
Application number	US-201615290648-A
Country	US
Kind code	B2
Filing date	Oct 11, 2016
Priority date	Oct 11, 2016
Publication date	Feb 14, 2023
Grant date	Feb 14, 2023

Abstract

Systems and methods to select a product title are described. The system identifies a set of item listings respectively describing items being offered for sale on a network-based marketplace. Each item listing includes a product identifier that matches and is not associated with a product title on the network-based marketplace. Each item listing also includes an item title. The system extracts feature values from the item listings and processes the feature values. The system evaluates the feature values to adopt a product title from an item title included in the set of item titles. The system generates a product user interface including the product title. Finally, the system communicates the product user interface, over a network, for display on a client machine. The product user interface includes the product title.

First claim

We claim:

1. A system for suggesting a title for an item listing of a network-based marketplace, comprising:
at least one hardware processor and instructions accessible on a computer-readable medium that, when executed, cause the at least one hardware processor to perform operations comprising:
identifying a plurality of item listings respectively describing items being offered for sale on the network-based marketplace, the identifying being based at least in part on each item listing including a product identifier that matches and is not associated with any product title on the network-based marketplace, the plurality of item listings including a corresponding plurality of item titles and a first item listing;
generating a plurality of item feature values from the plurality of item listings, each of the plurality of item feature values characterizing one item listing of the plurality of the item listings;
determining a plurality of product feature values from the plurality of item listings, each of the plurality of product feature values describing a comparison between one item listing of the plurality of item listings and multiple item listings of the plurality of item listings;
inputting one or more sets of item listings to a machine learning model;
receiving a training tag from an operator for the one or more sets of item listings to train the machine learning model to adopt at least one of the plurality of item titles for the one or more sets of item listings based on the product identifier, wherein the training tag classifies the at least one of the plurality of item titles as good for naming a product or bad for naming a product;
training the machine learning model based at least in part on inputting the one or more sets of item listings and the training tag;
adjusting, using the machine learning model, a plurality of weightings of the machine learning model for the plurality of product feature values;
weighting the plurality of item feature values and the plurality of product feature values in accordance with the adjustments to the plurality of weightings of the machine learning model;
analyzing the plurality of item feature values and the plurality of product feature values using the machine learning model to generate a plurality of item title scores corresponding to the plurality of item titles based at least in part on the weighting of the plurality of item feature values and the plurality of product feature values;
adopting, using the machine learning model, a product title from the at least one of the plurality of item titles based at least in part on the plurality of item title scores;
generating a product user interface suggesting the product title for the first item listing; and
communicating the product user interface to a client machine.

2. The system of claim 1 , wherein the operations further comprise enriching the product title, wherein the enriching includes reformatting a letter in the product title.

3. The system of claim 2 , wherein the reformatting the letter includes formatting the letter as uppercase.

4. The system of claim 1 , wherein the operations further comprise training the machine learning model based at least in part on a second plurality of items.

5. The system of claim 1 , wherein the product identifier is selected from a group of product identifiers comprising a global trade item number (GTIN), a manufacture part number (MPN), and a universal product code (UPC), and a European article number (EAN), a Japanese Article number (JAN), and an international standard book number (ISBN).

6. The system of claim 1 , wherein the operations further comprise generating the plurality of item title scores including title scores that are respectively associated with the item titles included in the plurality of item titles.

7. The system of claim 1 , where the plurality of product values includes a second product value including a Boolean describing whether a most frequently used token is included in a first item title.

8. The system of claim 1 , where the plurality of item feature values includes a first item feature value describing a range of title sizes.

9. The system of claim 1 , wherein the determining of the plurality of product feature values comprises determining a first product feature value of the plurality of product feature values as an average title length of the plurality of item listings, and wherein the determining of the plurality of item feature values comprises determining a first item feature value as a title length of the first item listing.

10. The system of claim 1 , wherein the determining of at least some of the plurality of product feature values characterizing item listing titles of at least two of the plurality of item listings exclude non-title values of the at least two of the plurality of item listings in the characterization.

11. The system of claim 1 , wherein the identifying the plurality of item listings based at least in part on each item listing including the product identifier that matches and is not associated with the product title on the network-based marketplace further comprises:
identifying the first item listing of the plurality of item listings that has not been initialized with a title or contains a null value as the title.

12. The system of claim 1 , wherein iteratively training of the machine learning model further comprises:
iteratively training, using a training set of item listings that respectively include item titles and a respective tag value assigned to each of the item titles, the machine learning model to differentiate between a first type of title and a second type of title.

13. A method for suggesting a title for an item listing of a network-based marketplace, comprising:
identifying a plurality of item listings respectively describing items being offered for sale on the network-based marketplace, the identifying being based at least in part on each item listing including a product identifier that matches and is not associated with any product title on the network-based marketplace, the plurality of item listings including a corresponding plurality of item titles and a first item listing;
generating, by at least one hardware processor, a plurality of item feature values from the plurality of item listings, each of the plurality of item feature values characterizing one item listing of the plurality of the item listings;
determining, by the at least one hardware processor, a plurality of product feature values from the plurality of item listings, each of the plurality of product feature values describing a comparison between one item listing of the plurality of item listings and multiple item listings of the plurality of item listings;
inputting one or more sets of item listings to a machine learning model;
receiving a training tag from an operator for the one or more sets of item listings to train the machine learning model to adopt at least one of the plurality of item titles for the one or more sets of item listings based on the product identifier, wherein the training tag classifies the at least one of the plurality of item titles as good for naming a product or bad for naming a product;
training the machine learning model based at least in part on inputting the one or more sets of item listings and the training tag;
adjusting, using the machine learning model, a plurality of weightings of the machine learning model for the plurality of product feature values;
weighting the plurality of item feature values and the plurality of product feature values in accordance with the adjustments to the plurality of weightings of the machine learning model;
analyzing

Assignees

• Ebay Inc

Inventors

• Dagan Arnon
• Zhicharevich Alexander

Classifications

• G06N20/00

  Machine learning · CPC title

• G06Q30/0641 Primary

  utilising user interfaces specially adapted for shopping · CPC title

• G06F16/9535 Primary

  Search customisation based on user profiles and personalisation · CPC title