A fork of the BANNER Named Entity Recognizer from Arizona State. This is based on a CVS snapshot from 11/27/2014.
- Add
BANNER_DATA
to your environment pointing to the absolute path to thebanner_data/
directory in this package. - Optionally, only if changes are made to the source code in this package, run:
sbt publish-local
.
The following changes were made to the original Banner code:
- Removed the following packages from
src
:test
,edu.umass.cs.mallet.projects
. They are not necessary for this project. - Moved the code from
src/
tosrc/main/java/
. - Moved all data resources (
dict/
,nlpdata/
,regex.txt
,banner.properties
) under the newbanner_data/
directory. Adjustedbanner.properties
to point to the new locations. - Trained a model using all BC2 training data, and saved it as
banner_data/banner_model.dat
. - Added the
BannerWrapper
class, which creates a Banner NER with default options, based on thebanner_data
directory. - Added the
BannerProperties.load(Properties properties)
method, which is needed by theBannerWrapper
class.
BANNER was created by Bob Leaman and is advised by Dr. Graciela Gonzalez, both members of the BioAI lab at Arizona State University. See the Banner project web page for more details: http://banner.sourceforge.net/. This project is distributed under the BANNER original license (see license.txt).
The packaging as a sbt project was implemented by Mihai Surdeanu and Marco Valenzuela.