Title:
WORD COLLECTION METHOD AND SYSTEM FOR USE IN WORD SEGMENTATION
Document Type and Number:
Japanese Patent JP2005251206
Kind Code:
A
Abstract:
To provide a method, a computer readable medium and a system which collect new words for addition to a lexicon for an agglutinative language.
In the method, a log of queries submitted to a search engine is obtained. The log of queries is sorted to obtain sorted queries. The sorted queries are then filtered using a plurality of heuristic criteria to obtain a candidate list of new words. Words from the candidate list of new words are then added to the lexicon.
Inventors:
OKUMURA KAORU
Application Number:
JP2005058934A
Publication Date:
September 15, 2005
Filing Date:
March 03, 2005
Export Citation:
Assignee:
MICROSOFT CORP
International Classes:
G06F17/30; G06F17/21; G06F17/28; G06F40/00; (IPC1-7): G06F17/30; G06F17/28
Domestic Patent References:
JPH04340163A | 1992-11-26 | |||
JPH09204437A | 1997-08-05 | |||
JPH04222055A | 1992-08-12 |
Attorney, Agent or Firm:
Yoshikazu Tani
Kazuo Abe
Kazuo Abe
Previous Patent: ASSISTED FORM INPUT
Next Patent: EVENT OWNERSHIP ALLOCATOR HAVING FAIL-OVER FOR MULTIPLE EVENT SERVER SYSTEMS
Next Patent: EVENT OWNERSHIP ALLOCATOR HAVING FAIL-OVER FOR MULTIPLE EVENT SERVER SYSTEMS