Browse Prior Art Database

Extraction of Viewpoint of Thesaurus using Corpora

IP.com Disclosure Number: IPCOM000116688D
Original Publication Date: 1995-Oct-01
Included in the Prior Art Database: 2005-Mar-31
Document File: 2 page(s) / 46K

Publishing Venue

IBM

Related People

Uramoto, N: AUTHOR

Abstract

Disclosed is a method for extracting "viewpoint" of classification of thesauruses. It enables the acquisition of deep knowledge for natural language processing systems, for example, system for machine translation and information retrieval. Thesaurus is a set of relationships between words, for example, "thing" and "food" (Fig. 1). A conventional thesaurus is created by human's intuition, so that the rule (viewpoint) of the classification of the thesaurus is not explicit. The method extracts the viewpoint by using the co-occurrence from large text corpora. This method has the following advantages. (1) it is possible to expand existing thesaurus. (2) it is possible to reflect the characteristics of domain-dependent text. (3) it is possible to subdivide the thesaurus. Fig.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 56% of the total text.

Extraction of Viewpoint of Thesaurus using Corpora

      Disclosed is a method for extracting "viewpoint" of
classification of thesauruses.  It enables the acquisition of deep
knowledge for natural language processing systems, for example,
system for machine translation and information retrieval.  Thesaurus
is a set of relationships between words, for example, "thing" and
"food" (Fig. 1).  A conventional thesaurus is created by human's
intuition, so that the rule (viewpoint) of the classification of the
thesaurus is not explicit.  The method extracts the viewpoint by
using the co-occurrence from large text corpora.  This method has the
following advantages.  (1) it is possible to expand existing
thesaurus.  (2) it is possible to reflect the characteristics of
domain-dependent text.  (3) it is possible to subdivide the
thesaurus.
  Fig. 1
  thing ---+-- food
        |
        +-- drink
        |
        +-- seasoning -- sugar
        |
        +-- shoes
        |
        +-- cloth --+-- wear
                     |
                     +-- kimono
  Fig. 2
  thing ---+-- food (eat, taste, enjoy)
        |
        +-- drink (drink, use)
        |
        +-- seasoning (add)
        |
        +-+ shoes (put on)
           |
           + cloth ---+-- wear
               (wear)  |
       ...