Skip to main content

Table 4 Statistics of relation extraction datasets

From: BioRel: towards large-scale biomedical relation extraction

Dataset

Word

Sentence

Entity

Relation

SemEval-2010

205k

10,717

21,434

9

ACE 2003-2004

297k

12,783

46,108

24

NYT

21,457k

695,059

17,816

54

BC5CDR

282k

11,089

29,271

1

BB3

34k

1394

2903

1

SeeDev

43k

1549

7082

22

GE4

134k

5130

13,012

5

i2b2 2010

91k

6310

8296

11

BioRel

26,166k

533,560

69,513

125