Browse Prior Art Database

Learning of Compound Word With Re-Cutting in Kana-to-Kanji Conversion

IP.com Disclosure Number: IPCOM000099968D
Original Publication Date: 1990-Mar-01
Included in the Prior Art Database: 2005-Mar-15
Document File: 1 page(s) / 33K

Publishing Venue

IBM

Related People

Ono, K: AUTHOR [+2]

Abstract

Disclosed is an algorithm for learning compound words. When cutting of the last conversion result is different from that of the first one, the algorithm learns the last compound word and grammar. After learning, the compound word is displayed at the first conversion.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 70% of the total text.

Learning of Compound Word With Re-Cutting in Kana-to-Kanji Conversion

       Disclosed is an algorithm for learning compound words.
When cutting of the last conversion result is different from that of
the first one, the algorithm learns the last compound word and
grammar.  After learning, the compound word is displayed at the first
conversion.

      Process-1:  Memorizes cutting boundary of the first conversion
result.

      Process-2:  Compares it with the last one (after next
conversion or re-cutting).  (If same, next operation.  If not same,
next process.)

      Process-3:  Checks if there is/are FUZOKUGO at left side of
cutting boundary.  If does not exist, next operation.  If exists,
next process.)
      Process-4:  Gets a grammar of right end word.
      Process-5:  Compounds (combines) a new word with every
                  cut word.
      Process-6:  Learns the new compound word.
      Example 1:  KOKUSAI JIN DOUHOU
      KOKUSAI/JIN/DOUHOU       p-1  Memorizes two boundaries
      KOKUSAIJIN/DOUHOU        p-2  Different boundaries
                               p-3  FUZOKUGO does not exist
                               p-4  Gets grammar, 'DOUHOU'
      KOKUSAIJINDOUHOU         p-5  Compounds (combines) them
                               p-6  Learns it
      Ex...