Geotagging unstructured text

Patent metadata

Field	Value
Publication number	US-9262438-B2
Application number	US-201313960119-A
Country	US
Kind code	B2
Filing date	Aug 6, 2013
Priority date	Aug 6, 2013
Publication date	Feb 16, 2016
Grant date	Feb 16, 2016

Abstract

Mechanisms are described to extract location information from unstructured text, comprising: building a language model from geo-tagged text; building a classifier for differentiating referred and physical location; given unstructured text, identifying referred location using the language model (that is, the location to which the unstructured text refers); given the unstructured text, identifying if referred location is also the physical location using the classifier; and predicting (that is, performing calculation(s) and/or estimation(s) of degree of confidence) of referred and physical location.

First claim

What is claimed is:

1. A method implemented in a computer system for extracting location information from unstructured text by utilizing a language model and a classifier, the method comprising:
obtaining, by a computer, the unstructured text;
identifying by the computer, via use of the language model and based upon the received unstructured text, a location referred to by the received unstructured text; and
determining by the computer, via use of the classifier, whether the location referred to by the received unstructured text is also a physical location from where the received unstructured text was sent;
wherein the language model is based upon a source of data that is distinct from the unstructured text.

2. The method of claim 1 , further comprising building, by the computer, the language model.

3. The method of claim 2 , wherein the language model is built based upon geo-tagged text.

4. The method of claim 2 , further comprising building, by the computer, a plurality of language models, each of the language models corresponding to a respective location.

5. The method of claim 1 , further comprising building, by the computer, the classifier.

6. The method of claim 5 , wherein the classifier is built based upon a training set of data.

7. The method of claim 1 , further comprising determining, by the computer, if the received unstructured text is location-neutral.

8. The method of claim 7 , wherein, if it is determined that the received unstructured text is location-neutral then the identifying the location referred to by the received unstructured text and the determining, via use of the classifier, whether the location referred to by the received unstructured text is also a physical location from where the received unstructured text was sent are not performed.

9. The method of claim 1 , wherein the identifying the location referred to by the received unstructured text comprises calculating, by the computer, a degree of confidence that the location referred to is correct.

10. The method of claim 1 , wherein the determining whether the location referred to by the received unstructured text is also the physical location from where the received unstructured text was sent comprises calculating, with the computer, a degree of confidence that the location referred to by the received unstructured text is also the physical location from where the received unstructured text was sent.

11. The method of claim 1 , further comprising outputting, by the computer, at least one of: (a) the location referred to by the received unstructured text; (b) the physical location from where the received unstructured text was sent; and (c) any combination thereof.

12. A computer readable storage medium, tangibly embodying a program of instructions executable by the computer for extracting location information from unstructured text by utilizing a language model and a classifier, the program of instructions, when executing, performing the following steps:
obtaining the unstructured text;
identifying, via use of the language model and based upon the received unstructured text, a location referred to by the received unstructured text; and
determining, via use of the classifier, whether the location referred to by the received unstructured text is also a physical location from where the received unstructured text was sent;
wherein the language model is based upon a source of data that is distinct from the unstructured text.

13. The computer readable storage medium of claim 12 , wherein the program of instructions, when executing, further performs building the language model.

14. The computer readable storage medium of claim 13 , wherein the language model is built based upon geo-tagged text.

15. The computer readable storage medium of claim 13 , wherein the program of instructions, when executing, further performs building a plurality of language models, each of the language models corresponding to a respective location.

16. The computer readable storage medium of claim 12 , wherein the program of instructions, when executing, further performs building the classifier.

17. The computer readable storage medium of claim 12 , wherein the program of instructions, when executing, further performs outputting at least one of: (a) the location referred to by the received unstructured text; (b) the physical location from where the received unstructured text was sent; and (c) any combination thereof.

18. A computer-implemented system for extracting location information from unstructured text by utilizing a language model and a classifier, the system comprising:
an input element configured to receive the unstructured text;
an identifying element configured to identify, via use of the language model and based upon the received unstructured text, a location referred to by the received unstructured text;
a determining element configured to determine, via use of the classifier, whether the location referred to by the received unstructured text is also a physical location from where the received unstructured text was sent; and
an output element configured to output the determination of whether the location referred to by the received unstructured text is also the physical location from where the received unstructured text was sent;
wherein the language model is based upon a source of data that is distinct from the unstructured text.

19. The system of claim 18 , further comprising a first building element configured to build the language model.

20. The system of claim 19 , wherein the language model is built based upon geo-tagged text.

21. The system of claim 19 , wherein the first building element is configured to build a plurality of language models, each of the language models corresponding to a respective location.

22. The system of claim 18 , further comprising a second building element configured to build the classifier.

23. The system of claim 18 , wherein the output element is further configured to output at least one of: (a) the location referred to by the received unstructured text; (b) the physical location from where the received unstructured text was sent; and (c) any combination thereof.

24. A method implemented in a computer system for extracting location information from unstructured text by utilizing a language model and a classifier, the method comprising:
building, by a computer, the language model;
building, by the computer, the classifier;
obtaining, by the computer, the unstructured text;
identifying by the computer, via use of the language model and based upon the received unstructured text, a location referred to by the received unstructured text;
determining by the computer, via use of the classifier, whether the location referred to by the received unstructured text is also a physical location from where the received unstructured text was sent; and
outputting, by the computer, at least one of: (a) the location referred to by the received unstructured text; (b) the physical location from where the received unstructured text was sent;
and (c) any combination thereof;
wherein the language model is based upon a source of data that is distinct from the unstructured text.

25. The method of claim 24 , wherein the language model is built based upon geo-tagged text.

Assignees

• IBM

Inventors

• Agrawal Dakshi
• Calo Seraphin B
• Ganti Raghu K
• Lee Kisung
• Srivatsa Mudhakar

Classifications

• G06F17/30705

  Physics · mapped topic

• G06F17/30241 Primary

  Physics · mapped topic

• G06F16/29 Primary

  Geographical information databases · CPC title

• G06F16/9537

  Spatial or temporal dependent retrieval, e.g. spatiotemporal queries · CPC title

• G06F16/35

  Clustering; Classification · CPC title