[Corpora-List] SVMTool++

Jesus Angel Gimenez Linares jgimenez at lsi.upc.edu
Mon Dec 6 15:59:01 CET 2004


Please apologize cross-postings


SVMTool++ v1.0
http://www.lsi.upc.edu/~nlp/SVMTool


We're glad to announce the release of the SVMTool++ v1.0, the
C++ version of the SVMTool, a simple, flexible, effective and efficient
part-of-speech tagger generator based on Support Vector Machines.

SVMTool++ is 10 times faster than the perl prototype, achieving a tagging
speed of more than 10,000 tokens per second. Results are also very
competitive,
achieving an accuracy of 97.2% for English on the Wall Street Journal
corpus.
It has also been successfully applied to Spanish and Catalan
exhibiting a similar
performance.

This version includes only the tagger and evaluator components (the
learner
is coming soon; in the meantime, models generated by SVMTool v1.2.1 perl
prototype are fully compatible with SVMTool++). Some of the additional
features (such as sentence-level tagging, or unsupervised models) are
not yet
supplied either.

The SVMTool is being developed a the TALP Research Center in Universitat
Politecnica de Catalunya (http://www.lsi.upc.es/~nlp) and is released
under LGPL.

please download the SVMTool++ v1.0 at http://www.lsi.upc.es/~nlp/SVMTool

contributions and comments are very welcome,

yours,

Jesus Gimenez & Lluis Marquez & Senen Moya








More information about the Corpora-archive mailing list