System and method for cross-document coreference

• US 6,438,543 B1
• Filed: 06/17/1999
• Issued: 08/20/2002
• Est. Priority Date: 06/17/1999
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Kazi, Zunaid H., Ravin, Yael
• Primary Examiner(s): Metjahic, Safet
• Assistant Examiner(s): CHEN, TE Y
• Application Number:
  US09/335,276
• Time in Patent Office: 1,160 Days
• Field of Search: 707/3-10, 707/104, 707/100-102, 707/1, 707/532, 707/536, 709/236, 716/5, 716/2, 345/762, 704/9-11, 704/2, 704/8
• US Class Current: 1/1
• CPC Class Codes: G06F 16/382 using citations hypermedia G06F16/94
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99937 Sorting
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing

Abstract

A method for coreferencing a plurality of documents includes the steps of providing a name list for names extracted from documents to be coreferenced upon entry of a query by a user, sorting the names of the list of names into mergable names and exclusive sets, comparing contexts of the mergable names against the exclusive sets to merge the mergable names to the exclusive sets exceeding a predetermined threshold to form an aggregated cross-document name list and referencing the aggregated cross-document name list to provide the user with coreferenced names across the plurality of documents which refer to a same entity in accordance with the query.

Claims

1. A method for coreferencing a plurality of documents comprising the steps of:

providing a name list for names extracted from documents to be coreferenced prior to or upon entry of a query by a user;

placing each name of the name list into canonical form;

sorting the names of the name list into mergable names and exclusive sets;

comparing contexts of the mergable names against the exclusive sets to merge the mergable names to the exclusive sets which exceed a predetermined threshold to form an aggregated cross-document name list; and

referencing the aggregated cross-document name list to provide the user with coreferenced names across the plurality of documents which refer to a same entity in accordance with the query.

2. The method as recited in claim 1, further comprising the step of extracting the name list from a collection of documents by employing a name extractor.

3. The method as recited in claim 1, further comprising the step of normalizing the name list to provide the names in the name list in a predetermined format.

4. The method as recited in claim 1, further comprising the step of splitting the names of the name list into component names based on evidence derived from the names.

5. The method as recited in claim 4, wherein the evidence derived from the names includes one of prefixes, suffixes, titles and information indicating one of a place, organization and a person.

6. The method as recited in claim 1, further comprising the step of merging identical names to reduce the name list.

7. The method as recited in claim 1, further comprising the step of sorting the name list into a person list and a place list.

8. The method as recited in claim 1, wherein the step of comparing contexts of the mergable names against the exclusive sets includes the steps of mapping all of the mergable names to each of the exclusive sets to provide matches above the predetermined threshold.

9. A program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform method steps for coreferencing a plurality of documents, the method steps comprising:

providing a name list for names extracted from documents to be coreferenced;

placing each name of the name list into canonical form;

sorting the names of the name list into mergable names and exclusive sets; and

comparing contexts of the mergable names against the exclusive sets to merge the mergable names to the exclusive sets which exceed a predetermined threshold to form an aggregated cross-document name list for the names in the name list.

10. The program storage device as recited in claim 9, further comprises the step of referencing the aggregated cross-document name list to provide a user with coreferenced names across the plurality of documents which refer to a same entity in accordance with a query input by the user.

11. The program storage device as recited in claim 9, further comprising the step of extracting the name list from a collection of documents by employing a name extractor.

12. The program storage device as recited in claim 9, further comprising the step of normalizing the name list to provide the names in the name list in a predetermined format.

13. The program storage device as recited in claim 9, further comprising the step of splitting the names of the name list into component names based on evidence derived from the names.

14. The program storage device as recited in claim 13, wherein the evidence derived from the names includes one of prefixes, suffixes, titles and information indicating one of a place, organization and a person.

15. The program storage device as recited in claim 9, further comprising the step of merging identical names to reduce the name list.

16. The program storage device as recited in claim 9, further comprising the step of sorting the name list into a person list and a place list.

17. The program storage device as recited in claim 9, wherein the step of comparing contexts of the mergable names against the exclusive sets includes the steps of mapping all of the mergable names to each of the exclusive sets to provide matches above the predetermined threshold.

18. A method for searching a plurality of documents for an entity having a plurality of variant names comprising the steps of:

providing a name list for names extracted from documents to be coreferenced prior to or upon entry of a search query by a user including a name of the entity;

placing each name of the name list into canonical form;

sorting the names of the name list into mergable names and exclusive sets;

comparing contexts of the mergable names against the exclusive sets to merge the mergable names to the exclusive sets which exceed a predetermined threshold to form an aggregated cross-document name list, the aggregated cross-document name list including a list of variant names for the entity; and

providing a list of documents to the user referencing the variant names and the name of the entity used for the search query.

19. The method as recited in claim 18, further comprising the step of extracting the name list from a collection of documents by employing a name extractor.

20. The method as recited in claim 18, further comprising the step of normalizing the name list to provide the names in the name list in a predetermined format.

21. The method as recited in claim 18, further comprising the step of splitting the names of the name list into component names based on evidence derived from the names.

22. The method as recited in claim 21, wherein the evidence derived from the names includes one of prefixes, suffixes, titles and information indicating one of a place, organization and a person.

23. The method as recited in claim 18, further comprising the step of merging identical names to reduce the name list.

24. The method as recited in claim 18, further comprising the step of sorting the name list into a person list and a place list.

25. The method as recited in claim 18, wherein the step of comparing contexts of the mergable names against the exclusive sets includes the steps of mapping all of the mergable names to each of the exclusive sets to provide matches above the predetermined threshold.