Parsing French using the MSTParser

home: Statistical dependency parsing of French

Université Paris Diderot - INRIA

Publications

Candito M.-H., Nivre J., Denis P. and Henestroza Anguiano E., 2010,
Benchmarking of Statistical Dependency Parsers for French, proceedings of COLING'2010 (poster session), Beijing, China

pdf

Parsing with MSTParser

Prerequisites

MSTParser version 0.4.3b, originally designed by Ryan McDonald
the MElt tagger (downloadable here) (Pascal Denis & Benoīt Sagot, Coupling an annotated corpus and a morphosyntactic lexicon for state-of-the-art POS tagging with less human effort. In Proceedings of PACLIC 2009, Hong Kong, China).
BONSAI v3.2 archive, to get preprocessing code and MST model (best MST model according to benchmark : uses predicted POS, predicted lemmas, predicted morpho features, and unsupervised word clusters). Note preprocessing code requires:
- perl and python >2.5
- python-cjson, to install with : python setup.py install
- recode >3.6
Set the MSTPARSER_DIR variable to your local path to MSTParser
Set the BONSAI variable to your local path to BONSAI v3.2

Parsing command

The following command will preprocess and parse a raw UTF-8 text file INFILE into INFILE.outmst :

$BONSAI/bin/bonsai_mst_parse_rawtext.sh [-n] INFILE

Use -n option if your text is already tokenized

Note : newlines in input text are systematically interpreted as sentence frontiers.