Skip to main content

Table 5 General statistics about agreement rates and concept assignments for the two corpora

From: Semantic annotation of biological concepts interplaying microbial cellular responses

 

Abstracts

Full-texts

 

F-scores

Final number of biological concepts

F-scores

Final number of biological concepts

dna

30.77%

25

13.22%

126

rna

81.82%

32

59.69%

119

gene

87.84%

73

91.78%

1175

protein

45.16%

35

42.15%

175

enzyme

70.18%

67

63.33%

388

transcription factor

20%

17

28.13%

47

compound

83.09%

188

63.90%

767

biochemical reaction

0%

-(*)

0%

-(*)

physiological state

46.63%

145

46.50%

403

laboratory technique

75.27%

58

38.34%

449

  1. The F-score columns refer to the F-score values achieved for the 130 documents after training and before post-processing; and the final number of biological concepts is calculated after post-processing.
  2. (*) This biological concept was not included in the final corpora. See the Post-processing sub-section for more details.