(13) Language Analysis Tools based on Unification-based Grammar

[Software Features]

A Japanese dependency analyzer based only on statistical information.
All the parameters used in the system have been learned from a
structurally analyzed Japanese corpus.  The current system uses
information related to the part-of-speech of the head of phrases,
distance between phrases, and occurence of punctuations.  The current
system just includes the learned parameters and does not include the
learning program.  The final version will include the learning module
and an interface to help the user compile parsed corpus.

[Required Environment]

*UNIX, gcc, Perl

*Japanese Morphological Analyzer "ChaSen version 1.0" is
 necessary, which is available at the following URL:

        URL: http://cactus.aist-nara.ac.jp/lab/nlt/chasen.html

[File Configuration]

*Data files
  Segmentation
  ./auto ./key ./zoku
  Dependency analysis
  ./Hstat

*Execution files
  ./unit        Segmentation(Perl)
  ./dep         Dependency analyzer(Perl)
  ./best        Best answer detector(C)
  ./ki          Tree writer(C)

[FTP]

• README
• Program and Documents in Japanese [327.3K]