Identifying location names within document text

• US 8,676,807 B2
• Filed: 04/22/2010
• Issued: 03/18/2014
• Est. Priority Date: 04/22/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Xiao, Rong, Yang, Jiangming, Zhang, Lei, Chen, Xingrong
• Primary Examiner(s): Chau, Dung K
• Application Number:
  US12/764,989
• Publication Number:
  US 20110264664A1
• Time in Patent Office: 1,426 Days
• Field of Search: None
• US Class Current: 707/741
• CPC Class Codes: G06F 16/31 Indexing; Data structures therefor; Storage structures

Abstract

Concepts and technologies are described herein for identifying location names within document text. Through an implementation of the concepts and technologies presented herein, functionality can be provided for identifying location names within articles, websites, travelogues, or other such documents. For instance, documents containing the names of cities, regions, countries, landmarks, or other locations may be associated with those locations. The location names may be unambiguously identified even when the location names may also have common word meanings that are not location associated or when the location name may be associated with more than one location.

Claims

1. A computer-implemented method for location name identification within document text, the computer-implemented method comprising performing computer-implemented operations for:

receiving a document;

detecting location names in the document;

recognizing location senses of the detected location names, wherein the location senses comprise a possible meaning or context for each location name;

organizing the recognized location senses together in a location sense hierarchy for the document comprising a location sense tree including observed location senses corresponding to location names detected within the document, and hidden location senses corresponding to location names not detected within the document but associated with the detected location names;

performing the detection of location names and the recognition of location senses iteratively;

determining a probability of each location name being an intended location, and determining a probability of each location name having a particular location sense;

generating an initial relevance score for each observed location sense based on the probability of each location name being the intended location, a number of times the observed location sense appears in the given document; and

generating a search index by propagating the relevance scores based on the initial relevance scores through the location sense tree such that the propagated relevance scores decay as they travel along branches of the location sense tree.

2. The computer-implemented method of claim 1, wherein the detection of location names and the recognition of location senses is initialized using statistics derived from a corpus of documents.

3. The computer-implemented method of claim 1, wherein the search index associates a relevance of location senses to the document.

4. The computer-implemented method of claim 1, further comprising applying heuristic clues to the detection of location names and the recognition of location senses.

5. The computer-implemented method of claim 1, wherein the detection of location names comprises disambiguation of location names from common words.

6. The computer-implemented method of claim 1, wherein the recognition of location senses comprises disambiguation of multiple location senses associated with a location name.

7. One of an optical, magnetic or solid state storage device having computer-readable instructions stored thereupon which, when executed by a computer, cause the computer to:

receive a document;

detect location names in the document;

recognize location senses of the detected location names, wherein the location senses comprise a possible meaning or context for each location name;

organize the recognized location senses together in a location sense hierarchy for the document comprising a location sense tree including observed location senses corresponding to location names detected within the document, and hidden location senses corresponding to location names not detected within the document but associated with the detected location names;

perform the detection of location names and the recognition of location senses iteratively;

determine a probability of each location name being an intended location, and determining a probability of each location name having a particular location sense;

generate an initial relevance score for each observed location sense based on the probability of each location name being the intended location, the probability of each location name having the particular location sense, and a number of times the observed location sense appears in the given document; and

generate a search index by propagating the relevance scores based on the initial relevance scores through the location sense tree such that the propagated relevance scores decay as they travel along branches of the location sense tree.

8. The one of an optical, magnetic or solid state storage device of claim 7, wherein the detection of location names and the recognition of location senses is initialized using statistics derived from a corpus of documents.

9. The one of an optical, magnetic or solid state storage device of claim 7, wherein the computer is further caused to apply heuristic clues to the detection of location names and the recognition of location senses.

10. The one of an optical, magnetic or solid state storage device of claim 7, wherein the detection of location names comprises disambiguation of location names from common words.

11. The one of an optical, magnetic or solid state storage device of claim 7, wherein the recognition of location senses comprises disambiguation of multiple location senses associated with a location name.

12. The one of an optical, magnetic or solid state storage device of claim 7, wherein the search index associates the relevance of location senses to the document.

13. A computer-implemented method for location name identification within document text, the computer-implemented method comprising performing computer-implemented operations for:

receiving a corpus;

receiving a document;

detecting location names, as identified from the corpus, in the document;

recognizing location senses, as identified from the corpus, of the detected location names, wherein the location senses comprise a possible meaning or context for each location name;

organizing the recognized location senses together in a location sense hierarchy for the document comprising a location sense tree including observed location senses corresponding to location names detected within the document, and hidden location senses corresponding to location names not detected within the document but associated with the detected location names;

performing the detection of location names and the recognition of location senses iteratively;

determining a probability of each location name being an intended location, and determining a probability of each location name having a particular location sense;

generating an initial relevance score for each observed location sense based on the probability of each location name being the intended location, the probability of each location name having the particular location sense, and a number of times the observed location sense appears in the given document; and

generating a search index by propagating the relevance scores based on the initial relevance scores through the location sense tree such that the propagated relevance scores decay as they travel along branches of the location sense tree.