CaboCha/南瓜: Yet Another Japanese Dependency Structure Analyzer

Cabocha is a SVM-based Japanese dependency parser. Since June, 2001, it achieved the best performance (89.29%) among the statistical parsing models for Japanese dependency parsing.

Features:

• Achieve high performance for dependency parsing based on SVMs
• Use the PKE(ACL2003) to speed up the SVM classifier
• Use deterministic cascaded chunking model and parse efficiently
• C/C++/Perl/Ruby

For more information, check Site: http://code.google.com/p/cabocha/