Evaluating Word Representation Features in Biomedical Named Entity Recognition Tasks

Table 1

Counts of different types of entities in two corpora used in this study.

Corpus	BioCreAtIvE II GM		JNLPBA
	Gene/protein	Total	Protein	DNA	RNA	Cell line	Cell type	Total
Training	18,265	18,265	30,269	9,534	951	3,830	6,718	51,301
Test	6,331	6,331	5,067	1,056	118	500	1,921	8,662