Skip to main content

Table 15 Concept normalization exact match results on the core evaluation annotation set of the 30 held-out documents compared to the baseline ConceptMapper approach

From: Concept recognition as a machine translation problem

Ontology

% OpenNMT class ID (%)

% ConceptMapper class ID (%)

% ConceptMapper FN Class ID (%)

% OpenNMT character (%)

% ConceptMapper character (%)

ChEBI

82*

55

41

94*

58

CL

72*

52

12

92*

77

GO_BP

82*

29

59

93*

36

GO_CC

81*

54

44

91*

55

GO_MF

98*

0

100

99*

0

MOP

95*

65

34

99*

66

NCBITaxon

87*

86

13

97*

87

PR

10

47*

26

76*

57

SO

97*

75

21

99*

78

UBERON

78*

64

34

95*

65

  1. We report both the percent exact match at the class ID level and the character level. We also report the percentage of false negatives (FN) for ConceptMapper (i.e., no class ID prediction for a given text mention). Note that for each ontology the better performance between OpenNMT and ConceptMapper is bolded with an asterisk* for both class ID and character levels
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us