Training Data for BioNLP'09 Shared Task on Event Extraction


- Contains 800 PubMed abstracts and their named entity and molecular event
  annotations.

- For format details, please refer to the shared task homepage.
  http://www-tsujii.is.s.u-tokyo.ac.jp/GENIA/SharedTask/


File Descriptions:

*.txt	files contain the target text.
	Each of these files consists of two lines, one for the title and one for the abstract.
*.a1	files contain standoff annotations for proteins.
	The participants will be provided with these files.
*.a2	files contain standoff annotations for other named entities and events.
	For training, the participants will be provided with these files.
	For test, the participants have to create this annotation.

History:

 3 July 2009, released to public.

10 February 2009, 2nd revision released.
	- Cause arguments from Phosphorylation events have been removed: 7 cases.
	- Duplicated events have been removed: 127 cases.
	- Duplicated Equiv entities have been removed: 8 cases.
	- Redundant Entity annotations have been removed: 20 cases.

 1 February 2009, 1st revision released.

21 January  2009, The first version released.


Authors:

BioNLP'09 Shared Task organizers,
bionlpsharedtask@gmail.com
http://www-tsujii.is.s.u-tokyo.ac.jp/GENIA/SharedTask/