Automatic lot classification

Patent metadata

Field	Value
Publication number	US-12001471-B2
Application number	US-202117245406-A
Country	US
Kind code	B2
Filing date	Apr 30, 2021
Priority date	Mar 8, 2018
Publication date	Jun 4, 2024
Grant date	Jun 4, 2024

Abstract

Methods, systems, and media for lot classification are disclosed. In one example, a classification system for identifying lot listings receives a description for a listing in a publication system, identifies a string in the listing, identifies a quantity word or digit in the string, and converts an identified quantity word into digit form. A normalized string is tokenized to produce tokens, the tokenizing of the normalized string including splitting the normalized string into a series of substrings using a sequence of delimiters. For each substring, an additional split is performed by separating any digit from any other adjacent character, unless that character is another digit, and maintaining an internal character order of each split substring to produce a flattened list of tokenized tokens.

First claim

What is claimed is:

1. A computer implemented method comprising:
receiving an input comprising a string indicating a plurality of items are included in a single listing for publication by a publication system;
parsing, by one or more processors, the string to identify a token that comprises a quantity word or a digit, the token being formed by:
identifying the quantity word in the string;
converting the identified quantity word into digit form;
normalizing the string to lowercase characters;
splitting the normalized string into a series of substrings using a sequence of delimiters; and
splitting the series of substrings by separating a character from an adjacent character based on a difference between the character and the adjacent character;

determining, by a machine learning model, a probability that the token is indicative of a quantity of the plurality of items;
classifying, by the one or more processors, the single listing as a lot listing corresponding to the plurality of items based at least in part on the probability that the token is indicative of the quantity of the plurality of items included in the lot listing; and
causing display of the single listing as the lot listing based at least in part on the classifying.

2. The method of claim 1 , wherein classifying the single listing as the lot listing is based at least in part on a position of the token in the string relative to one or more other tokens in the string.

3. The method of claim 2 , wherein determining the probability further comprises determining the probability that the quantity word or digit corresponds to the lot listing based at least in part on the position of the token in the string relative to the one or more other tokens in the string, wherein classifying the single listing as the lot listing is based at least in part on the probability.

4. The method of claim 1 , wherein the input is a first input, the string is a first string, the single listing is a first listing, the quantity word or digit is a first quantity word or digit, the method further comprising:
receiving a second input comprising a second string indicating a second plurality of items are included in a second listing for publication by the publication system, wherein the second input is different than the first input, the second string is different than the first string, and the second listing is different than the first listing;
parsing, by the one or more processors, the second string to identify a second token;
determining, by the machine learning model, a probability that the second token comprises a second quantity word or digit; and
classifying, by the one or more processors, the second listing as a second type of listing based at least in part on the probability that the second token comprises the second quantity word or digit.

5. The method of claim 4 , wherein classifying the second listing as the second type of listing further comprises classifying, by the one or more processors, the second listing as the second type of listing based at least in part on a position of the second token in the second string.

6. The method of claim 5 , further comprising causing display of the second listing as the second type of listing based at least in part on classifying the second listing as the second type of listing.

7. A system, comprising:
a processor; and
a memory device storing instructions which, when executed by the processor, causes the system to perform operations comprising:
receiving an input comprising a string indicating a plurality of items are included in a single listing for publication by a publication system;
parsing the string to identify a token that comprises a quantity word or a digit, the token being formed by:
identifying the quantity word in the string;
converting the identified quantity word into digit form;
normalizing the string to lowercase characters;
splitting the normalized string into a series of substrings using a sequence of delimiters; and
splitting the series of substrings by separating a character from an adjacent character based on a difference between the character and the adjacent character;

determining, by a machine learning model, a probability that the token is indicative of a quantity of the plurality of items;
classifying the single listing as a lot listing corresponding to the plurality of items based at least in part on the probability that the token is indicative of the quantity of the plurality of items included in the lot listing; and
causing display of the single listing as the lot listing based at least in part on the classifying.

8. The system of claim 7 , wherein classifying the single listing as the lot listing is based at least in part on a position of the token in the string relative to one or more other tokens in the string.

9. The system of claim 8 , wherein the processor, when executing the instructions to determine the probability, causes the system to perform operations comprising determining the probability that the quantity word or digit corresponds to the lot listing based at least in part on the position of the token in the string relative to the one or more other tokens in the string, wherein classifying the single listing as the lot listing is based at least in part on the probability.

10. The system of claim 7 , wherein the input is a first input, the string is a first string, the single listing is a first listing, the quantity word or digit is a first quantity word or digit, and the processor, when executing the instructions, causes the system to perform operations comprising:
receiving a second input comprising a second string indicating a second plurality of items are included in a second listing for publication by the publication system, wherein the second input is different than the first input, the second string is different than the first string, and the second listing is different than the first listing;
parsing the second string to identify a second token;
determining, by the machine learning model, a probability that the second token comprises a second quantity word or digit; and
classifying the second listing as a second type of listing based at least in part on the probability that the second token comprises the second quantity word or digit.

11. The system of claim 10 , wherein the processor, when executing the instructions, causes the system to perform operations comprising classifying the second listing as the second type of listing based at least in part on a position of the second token in the second string.

12. The system of claim 11 , wherein the processor, when executing the instructions, causes the system to perform operations comprising causing display of the second listing as the second type of listing based at least in part on classifying the second listing as the second type of listing.

13. A non-transitory computer-readable medium comprising instructions which, when read by a machine, cause the machine to perform operations comprising:
receiving an input comprising a string indicating a plurality of items are included in a single listing for publication by a publication system;
parsing the string to identify a token that comprises a quantity word or a digit, the token being formed by:
identifying the quantity word in the string;
converting the identified quantity word into digit form;
normalizing the string to lowercase characters;
splitting the normalized string into a series of substrings using a sequence of delimiters; and
splitting the series of substrings by separating a character from an adjacent character based on a difference between the character and the adjacent character;

Assignees

• Ebay Inc

Inventors

• Hurwitz Daniel Lee
• Guy Ido

Classifications

• G06F16/358 Primary

  Browsing; Visualisation therefor · CPC title

• G06F16/35 Primary

  Clustering; Classification · CPC title

• G06N20/00

  Machine learning · CPC title

• G06F40/279

  Recognition of textual entities · CPC title

• G06F40/163

  Handling of whitespace · CPC title