APPARATUS, SYSTEM AND METHOD FOR STRING DISAMBIGUATION AND ENTITY RANKING

• US 20160357738A1
• Filed: 09/30/2015
• Published: 12/08/2016
• Est. Priority Date: 06/07/2015
• Status: Active Grant

• Current Assignee: Apple Inc.
• Original Assignee: Apple Inc.
• Inventors: Langmead, Gregory C., Hughes, Dominic J., Gingrich, Mark A., Jammalamadaka, Ravi Chandra, Beaver, Donald R., Liu, Shizhu, Bhirud, Pushkaraj, Ali, Kamal M., Silva, Daniel Ribeiro, Murrett, Martin J.
• Granted Patent: US 10,146,775 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: G06F 16/374 Thesaurus
• CPC Class Codes: G06F 16/93 Document management systems
• CPC Class Codes: G06F 40/134 Hyperlinking
• CPC Class Codes: G06F 40/295 Named entity recognition
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

Aspects of the present disclosure involve a mobile or computer reader application that obtains articles or other computer files from a central database and displays the articles to a user of the device. The reader application may be customizable around one or more characteristics of the user of the device. In one embodiment, the type and number of articles provided to the device and displayed in the reader application may be based on the determination of a category or type of usage of the application is performed by the user. Further, the determination of the use of the reader application on the device is performed by and contained within the device such that usage information is not shared with overall article providing system. In another embodiment, the article providing system and/or device may determine recommendations to provide to a user of the reading application. These recommendations may be based on one or more selected interests or topics of the user of the reading application.

Claims

1. A method for classifying an article, the method comprising:

spotting an article to identify one or more entities, the one or more entities comprising strings of text within the article;

determining the location within the article of the occurrences of at least one entity of the one or more entities;

assigning a weighted score to each of the occurrences of the at least one entity of the one or more entities, the weighted score based at least on the location within the article of each of the occurrences;

summing the weighted score for each of the occurrences of the at least one entity of the one or more entities to calculate an entity importance score;

assigning the entity importance score to the at least one entity of the one or more entities; and

storing the entity importance score and an identifier of the at least one entity with the article in an article database.

2. The method of claim 1 wherein the location within the article is at least one of an article title, an article description, an article uniform resource locator, an article content portion, or a first sentence of the article.

3. The method of claim 2 further comprising:

adjusting the weighted score of at least one occurrence of the at least one entity of the one or more entities by a decay curve score value when the location within the article is the article content portion.

4. The method of claim 2 wherein the weighted score assigned to each occurrence of the at least one entity of the one or more entities in the title is more than the weighted score assigned to each occurrence of the at least one entity of the one or more entities in the article description.

5. The method of claim 1 further comprising:

reducing the weighted score for the at least one entity if the location within the article of the at least one entity is only in a beginning portion of the article.

6. The method of claim 1 further comprising:

calculating a percentage of sentence occurrences of the at least one entity in the article by dividing the total number of text sentences in the article in which the at least one entity occurs by the total number of text sentences within the article; and

adjusting the entity importance score for the at least one entity of the one or more entities based on the calculated percentage of sentence occurrences of the at least one entity in the article.

7. The method of claim 1 further comprising:

accessing the knowledge database of related entities to determine a synonymous relationship between the at least one entity of the article to a second entity of the one or more entities of the article.

8. The method of claim 7 further comprising:

combining the weighted score of the at least one entity of the article and a weighted score of the second entity of the one or more entities of the article to calculate a combined synonymous weighted score.

9. The method of claim 1 further comprising:

accessing a knowledge database of related entities to determine a taxonomy relationship between the at least one entity of the article to at least another entity of the one or more entities of the article.

10. The method of claim 9 wherein the taxonomy relationship between the at least one entity of the article to at least another entity of the one or more entities of the article comprises a hyperlink from the at least one entity of the article to at least another entity of the one or more entities of the article in the knowledge database.

11. A system for classifying a collection of computer files, the system comprising:

an article database storing a plurality of article-based computer files;

a spotting device in communication with the article database to retrieve a first article of the plurality of article-based computer files from the article database and identify one or more entities within a first article of the plurality of article-based computer files, the one or more entities comprising strings of text within the first article; and

an entity ranker device executing one or more computer-executable instructions to;

determine the location within the article of the occurrences of at least one entity of the one or more entities;

assign a weighted score to each of the occurrences of the at least one entity of the one or more entities, the weighted score based at least on the location within the first article of each of the occurrences;

sum the weighted score for each of the occurrences of the at least one entity of the one or more entities to calculate an entity importance score;

assign the entity importance score to the at least one entity of the one or more entities; and

store the entity importance score and an identifier of the at least one entity with the first article in the article database.

12. The system of claim 11 wherein the location within the article is at least one of an article title, an article description, an article uniform resource locator, an article content portion, or a first sentence of the article.

13. The system of claim 12 wherein the entity ranker further adjusts the weighted score of at least one occurrence of the at least one entity of the one or more entities by a decay curve score value when the location within the first article is the article content portion.

14. The system of claim 12 wherein the weighted score assigned to each occurrence of the at least one entity of the one or more entities in the title is more than the weighted score assigned to each occurrence of the at least one entity of the one or more entities in the article description.

15. The system of claim 11 wherein the entity ranker further calculates a percentage of sentence occurrences of the at least one entity in the first article by dividing the total number of text sentences in the first article in which the at least one entity occurs by the total number of text sentences within the first article and adjusts the entity importance score for the at least one entity of the one or more entities based on the calculated percentage of sentence occurrences of the at least one entity in the first article.

16. The system of claim 11 wherein the entity ranker further accesses the knowledge database of related entities to determine a synonymous relationship between the at least one entity of the article to a second entity of the one or more entities of the article.

17. The system of claim 16 wherein the entity ranker further combines the weighted score of the at least one entity of the article and a weighted score of the second entity of the one or more entities of the article to calculate a combined synonymous weighted score.

18. A non-transitory computer-readable medium storing one or more executable instructions to perform a computer-implemented method, the one or more executable instructions to:

spotting an article to identify one or more entities comprising strings of text within the article and the location within the article of the occurrences of at least one entity of the one or more entities;

assigning a weighted score to each of the occurrences of the at least one entity of the one or more entities, the weighted score based at least on the location within the article of each of the occurrences;

summing the weighted score for each of the occurrences of the at least one entity of the one or more entities to calculate an entity importance score;

assigning the entity importance score to the at least one entity of the one or more entities; and

storing the entity importance score and an identifier of the at least one entity with the article in an article database.

19. The non-transitory computer-readable medium of claim 18 wherein the location within the article is at least one of an article title, an article description, an article uniform resource locator, an article content portion, or a first sentence of the article.

20. The non-transitory computer-readable medium of claim 19 wherein the computer-implemented method further comprises:

adjusting the weighted score of at least one occurrence of the at least one entity of the one or more entities by a decay curve score value when the location within the article is the article content portion.