GitHub - timfeu/berkeleycoref-bansalklein: Extractor for Bansal & Klein features used in the Berkeley Coreference System modified with distributional features

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
code		code
config		config
lib		lib
models		models
sample_data		sample_data
LICENSE		LICENSE
README		README
build.xml		build.xml

Repository files navigation

This software extracts the features described in the following paper:

"Coreference semantics from web features"
Mohit Bansal and Dan Klein. Proceedings of ACL 2012.

to be used with the Berkeley Coreference Resolution System described in:

"Easy Victories and Uphill Battles in Coreference Resolution"
Greg Durrett and Dan Klein. EMNLP 2013.

A modified version of the Berkeley system supporting these features can be found at https://github.com/timfeu/berkeleycoref-thesaurus

==========
Build
==========

Use Apache Ant (http://ant.apache.org) to build a fat jar:

ant jar

==========
Run
==========

java -jar build/jar/runnable_wrapper_and_data.jar mode [options]

Where mode is one of the following:

generalCoOccurrence
hearst
entity
pronoun
cluster

===========
License
===========

This software is licensed under GPLv3. It includes the Stanford POS tagger
released under GPLv2+: http://nlp.stanford.edu/software/tagger.shtml