Skip to main content

Table 22 Results of constituent parsers using their distributed biomedical models on the CRAFT release set; labeled bracket precision (LB-P), recall (LB-R) and F-score (LB-F)

From: A corpus of full-text journal articles is a robust evaluation tool for revealing differences in performance of biomedical natural language processing tools

Parser

LB-P

LB-R

LB-F

unevaluated count

Charniak-Johnson

51.23

55.99

53.50

0

Charniak-Lease

53.28

57.43

55.28

8

Mogura

47.55

56.27

51.54

105

Stanford 1.6.6

59.49

61.81

60.63

10