Table 1.

Statistics of the NCBI, GM, and CDR corpora

Corpus	Entity	Unit	Training	Develop	Test	Total (Unit)
NCBI	Disease	Abstracts	592	100	100	792 (abstracts)
GM	Gene	Sentences	15000	-	5000	20000 (sentences)
CDR	Disease, Chemicals	Abstracts	500	500	500	1500 (abstracts)