Entity name recognition

• US 7,792,837 B1
• Filed: 11/14/2007
• Issued: 09/07/2010
• Est. Priority Date: 11/14/2007
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Zhao, Shubin
• Primary Examiner(s): NGUYEN, CAM LINH T
• Application Number:
  US11/940,177
• Time in Patent Office: 1,028 Days
• Field of Search: 707/1, 707/3, 707/4, 707/6, 707/7, 707/10, 707/100, 707/104.1, 707/200, 707/204
• US Class Current: 707/738
• CPC Class Codes: G06F 16/313 Selection or weighting of terms for indexing

Abstract

A system, method, and computer program product for recognizing entity names from a plurality of documents. Embodiments of the methods comprise selecting a selection of documents from a plurality of documents, the selection of documents sharing a common pattern in their titles. The method further comprises determining a name candidate for each document in the selection by applying the common pattern to the title of the document, and matching the name candidates with a collection of entity names (the white list). Responsive to determining a match between the name candidates and the entity names in the white list, the method determines that the name candidates are valid entity names. In one embodiment, the name candidates are added to the white list after being determined to be valid entity names.

Claims

1. A method for recognizing entity names from a plurality of web documents, comprising:

at a server having one or more processors and memory storing one or more programs executed by the one or more processors,

identifying a common pattern in titles of the plurality of web documents;

selecting a selection of documents from the plurality of web documents, a title of each document in the selection sharing the common pattern;

for each document in the selection, generating an entity name candidate from the title of the document in accordance with the common pattern;

matching the entity name candidates of the selection of documents with a collection of entity names;

determining that the entity name candidates of the documents in the selection are valid entity names in accordance with a result of the matching; and

updating the collection of entity names in accordance with a result of the determining.

2. The method of claim 1, wherein selecting the selection of documents from the plurality of web documents further comprises:

selecting the selection of documents from the plurality of web documents, each document in the selection being from a same domain.

3. The method of claim 1, wherein the common pattern comprises a prefix section, a body section, and a suffix section.

4. The method of claim 3, wherein generating the entity name candidate by applying the common pattern to the title of the document further comprises:

generating the entity name candidate by removing the prefix section and the suffix section of the common pattern from the title of the document.

5. The method of claim 1, wherein matching the entity name candidates of the selection of documents with the collection of entity names further comprises determining whether at least one of the following conditions is satisfied:

a number of the entity name candidates matching the entity names in the collection exceeds a threshold;

a number of the entity names in the collection matching the entity name candidates exceeds a threshold; and

a portion of the entity name candidates matching with entity names in the collection exceeds a threshold.

6. The method of claim 1, wherein an entity name candidate and an entity name in the collection of entity names are considered matched when they are identical.

7. The method of claim 1, wherein an entity name candidate and an entity name in the collection of entity names are considered matched when their edit distance is smaller than a threshold.

8. The method of claim 1, wherein updating the collection of entity names comprises adding the entity name candidates of the documents in the selection to the collection of entity names in accordance with the result of the matching.

9. The method of claim 1, further comprising:

matching the entity name candidates of the selection of documents with a collection of texts; and

determining that the entity name candidates of the documents in the selection are invalid entity names in accordance with the result of the matching.

10. The method of claim 1, wherein generating the entity name candidate from the title of the document includes normalizing the titles of the documents in the selection to generate normalized titles.

11. The method of claim 1, wherein matching the entity name candidates of the selection of documents with the collection of entity names includes normalizing the entity name candidates of the selection of documents to generate normalized entity name candidates.

12. A method for recognizing entity names from a plurality of web documents, comprising:

at a server having one or more processors and memory storing one or more programs executed by the one or more processors,

(a) identifying common patterns in titles of the plurality of web documents;

(b) selecting multiple selections of documents from the plurality of web documents, titles of documents in each selection sharing a respective common pattern associated with the selection of documents;

(c) for each of the multiple selections of documents, generating a selection of entity name candidates by applying the respective common pattern to the titles of the documents in the selection of documents; and

(d) in response to matching a collection of entity names and a selection of entity name candidates, performing steps comprising;

(i) determining that the selection of name candidates are valid entity names,

(ii) updating the collection of entity names in accordance with a result of the determining,

(iii) removing a selection of documents associated with the selection of entity name candidates from the multiple selections of documents, and

(iv) repeating step (d).

13. A system for recognizing entity names from a plurality of web documents, the system comprising:

a processor for executing computer program code; and

a subsystem executable by the processor, the subsystem including computer program code for performing a method comprising;

identifying a common pattern in titles of the plurality of web documents;

selecting a selection of documents from the plurality of web documents, a title of each document in the selection sharing the common pattern;

for each document in the selection, generating an entity name candidate from the title of the document in accordance with the common pattern;

matching the entity name candidates of the selection of documents with a collection of entity names;

determining that the name candidates of the documents in the selection are valid entity names in accordance with a result of the matching; and

updating the collection of entity names in accordance with a result of the determining.

14. The system of claim 13, wherein selecting the selection of documents from the plurality of web documents further comprises:

selecting the selection of documents from the plurality of web documents, each document in the selection being from a same domain.

15. The system of claim 13, wherein:

the common pattern comprises a prefix section, a body section, and a suffix section;

generating the entity name candidate from the title of the document further comprises generating the entity name candidate by removing the prefix section and the suffix section of the common pattern from the title of the document.

16. The system of claim 13, wherein updating the collection of entity names comprises adding the entity name candidates of the documents in the selection to the collection of entity names in accordance with the result of the matching.

17. A computer readable storage medium storing one or more programs for execution by one or more processors of a computer system, the one or more programs including:

instructions for identifying a common pattern in titles of a plurality of web documents;

instructions for selecting a selection of documents from the plurality of web documents, a title of each document in the selection sharing the common pattern;

instructions for generating an entity name candidate for each document in the selection from the title of the document in accordance with the common pattern;

instructions for matching the entity name candidates of the selection of documents with a collection of entity names;

instructions for determining that the name candidates of the documents in the selection are valid entity names in accordance with a result of the matching; and

instructions for updating the collection of entity names in accordance with a result of the determining.

18. The computer readable storage medium of claim 17, wherein the instructions for selecting the selection of documents from the plurality of web documents further comprise:

instructions for selecting the selection of documents from the plurality of web documents, each document in the selection being from a same domain.

19. The computer readable storage medium of claim 17, wherein:

the common pattern comprises a prefix section, a body section, and a suffix section; and

the instructions for generating the entity name candidate for each document in the selection from the title of the document further comprises instructions for generating the entity name candidate by removing the prefix section and the suffix section of the common pattern from the title of the document.

20. The computer program product of claim 17, wherein the instructions for updating the collection of entity names includes instructions for adding the entity name candidates of the documents in the selection to the collection of entity names in accordance with the result of the matching.