Skip to content

The program used in the paper 'Gut, Besser, Chunker – Selecting the best models for text chunking with voting' by Balázs Indig and István Endrédy

License

Notifications You must be signed in to change notification settings

sfu-natlang/gut-besser-chunker

 
 

Repository files navigation

Gut, Besser, Chunker

The program used in the paper 'Gut, Besser, Chunker – Selecting the best models for text chunking with voting' by Balázs Indig and István Endrédy

Prerequisites

  • data folder: train.txt and test.txt (The CoNLL-2000 data)
  • CRFsuite installed and (crfsuite and chunking.py in path)
  • TnT installed and in path
  • All git submodules inited and updated (./postClone.sh)
  • Python3, JRE 1.8, NLTK 3 installed
  • PurePOS compiled and the path is set in IOB_tagger_purepos.sh

Usage

  • Run ./start.sh and wait...
  • See .sh and and .py files for documentation...
  • PurePOS as an IOB tagger can be invoked separately from IOB_tagger_purepos.sh (You must set PATHTOPUREPOS variable first!).

Warning

  • Converters may contain bugs. Please check intermediate data files!
  • HunTag3 has very slow learning-time on lexicalised input
  • The whole process takes much time!

License

The repository contains many 3rd-party code, that has its own license. This code is made available under the GNU Lesser General Public License v3.0.

Reference

If you use this program please cite:

	@inproceedings{gut-besser-chunker,
	  author    = {{I}ndig, {B}al\'azs and {I}stv\'an, {E}ndr\'edy},
	  title     = {{Gut, Besser, Chunker} -- {S}electing the best models for text chunking with voting},
	  booktitle = {Computational Linguistics and Intelligent Text Processing - 17th International Conference, CICLing 2016, Konya, Turkey, April 3-9, 2016},
	  year      = {2016},
	  publisher  = {Springer} 
	}

About

The program used in the paper 'Gut, Besser, Chunker – Selecting the best models for text chunking with voting' by Balázs Indig and István Endrédy

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 55.6%
  • Perl 30.5%
  • Shell 12.4%
  • Java 1.5%