Method or system for semantic categorization

• US 9,305,103 B2
• Filed: 07/03/2012
• Issued: 04/05/2016
• Est. Priority Date: 07/03/2012
• Status: Active Grant

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Suprasadachandranpilliai, Syama Prasad
• Primary Examiner(s): Truong, Dennis
• Application Number:
  US13/541,566
• Publication Number:
  US 20140012854A1
• Time in Patent Office: 1,372 Days
• Field of Search: 707/737, 707/740, 707/754, 707/759, 707/722, 707/728, 707/730, 707/767, 707/768
• US Class Current: 1/1
• CPC Class Codes: G06F 16/954 Navigation, e.g. using categorised browsing

Abstract

Methods and/or systems are provided that may be utilized to rank categories of an entity based at least in part on relevance.

Claims

1. A method comprising:

determining, via one or more processors, an entity based at least in part on a submitted search query;

processing, via the one or more processors, a plurality of lists and a plurality of categories in which the entity has been grouped in at least one content source prior to receiving the submitted search query; and

determining, via the one or more processors, a particular category, from the plurality of categories, to present, at least partially in response to the submitted search query, for the entity based at least in part on semantic categorization, wherein the semantic categorization is based at least in part on a sum of products over the plurality of lists, in which for a particular list of the plurality of lists, and a product comprises a product of a popularity value of the entity with respect to the particular list and a probability value of the particular list being within the particular category, wherein the popularity value is based, at least on part, on a number of the plurality of lists in which the entity has been grouped.

2. The method of claim 1, and further comprising:

serving the particular category as an alternate search query to the submitted query or as an alternate for a portion thereof.

3. The method of claim 1, wherein the determining a particular category for an entity comprises:

determining respective category scores for more than one category associated with the entity; and

selecting at least one category of the more than one category based at least in part on the respective category scores.

4. The method of claim 3, wherein the determining the respective category scores comprises, for the particular category over the plurality of lists:

measuring interest for the entity of a list from the plurality of lists;

measuring likelihood of the list from the plurality of lists being associated with the particular category of the more than one category; and

calculating the category score for the particular category over the plurality of lists based at least in part on the interest and likelihood measurements.

5. The method of claim 4, wherein the measured likelihood comprises a binary value of one or zero;

and wherein the calculating the category score comprises;

computing a product of the measured interest and the measured likelihood for a particular list of the plurality of lists; and

summing the computed products over the plurality of lists.

6. The method of claim 4, further comprising acquiring, from one or more content sources, the entity, the plurality of lists, and the more than one category.

7. The method of claim 6, wherein the acquiring the entity, the plurality of lists, and the more than one category further comprises acquiring one or more titles and/or one or more descriptions.

8. The method of claim 7, wherein the acquiring further comprises pruning punctuation from the more the one category, the one or more titles and/or the one or more descriptions.

9. The method of claim 7, wherein the acquiring further comprising applying natural language processing (NLP) to the more the one category, the one or more titles and/or the one or more descriptions to determine the entity, the plurality of lists, and the more than one category.

20. The method of claim 1, wherein the particular list of the plurality of lists are included within one or more of the plurality of categories in the at least one content source.

10. A method comprising:

processing, by one or more processors, one or more signals comprising a particular category, determined from a plurality of categories, to present, at least partially in response to the submitted search query, for an entity based at least in part on semantic categorization, wherein the semantic categorization is based at least in part on a sum of products over the plurality of lists, in which for a particular list of a the plurality of lists in which an entity has been grouped in at least one content source prior to receiving a submitted search query, and a product comprises a product of a popularity value of the entity with respect to a particular list and a probability value of the particular list being within the particular category;

wherein individual ones of the plurality of categories being independent of any submitted search queries and based, at least in part, on the plurality of lists in which the entity has been grouped in the at least one content source, and the popularity value is based, at least on part, on a number of the plurality of lists in which the entity has been grouped.

11. The method of claim 10, wherein the particular category comprises an alternate search query to the submitted query or as an alternate for a portion thereof.

12. The method of claim 10, wherein the determining the particular category for an entity comprises:

determining respective category scores for more than one category associated with the entity; and

selecting at least one category of the more than one category based at least in part on the respective category scores.

13. The method of claim 12, wherein the determining the respective category scores comprises, for the particular category over the plurality of lists:

measuring interest for the entity of a list from the plurality of lists;

measuring likelihood of the list from the plurality of lists being associated with the particular category of the more than one category; and

calculating the category score for the particular category over the plurality of lists based at least in part on the interest and likelihood measurements.

14. The method of claim 13, wherein the measured likelihood comprises a binary value of one or zero;

and wherein the calculating the category score comprises;

computing a product of the measured interest and the measured likelihood for a particular list of the plurality of lists; and

summing the computed products over the plurality of lists.

15. An apparatus comprising:

a computing platform;

said computing platform to;

determine an entity to be based at least in part on a search query to be submitted;

process a plurality of lists and a plurality of categories in which the entity has been grouped in at least one content source prior to receipt of the search query; and

determine a particular category, from a plurality of categories, to present, at least partially in response to the submitted search query, for the entity to be based at least in part on semantic categorization, wherein the semantic categorization is to be based at least in part on a sum of products over the plurality of lists, in which for a particular list of the plurality of lists, a product is to comprise a product of a popularity value of the entity with respect to the particular list and a probability value of the particular list being within the particular category, and wherein the popularity value is to be based, at least on part, on a number of the plurality of lists in which the entity has been grouped.

16. The apparatus of claim 15, said computing platform to further serve the category as an alternate search query to the submitted query or as an alternate for a portion thereof.

17. The apparatus of claim 15, wherein the computing platform further to determine respective category scores for more than one category associated with the entity;

and to select at least one category of the more than one category based at least in part on the respective category scores to be determined.

18. The apparatus of claim 17, wherein the computing platform to further, for the particular category over the plurality of lists:

measure interest for the entity of a list from the plurality of lists;

measure likelihood of the list from the plurality of lists to be associated with the particular category of the more than one category; and

calculate the category score for the particular category over the plurality of lists to be based at least in part on the interest and likelihood measurements.

19. The apparatus of claim 18, the computing platform to further compute a product of the interest to be measured and the likelihood to be measured for a particular list of the plurality of lists;

and to sum the products to be computed over the plurality of lists.