Skip to main content

Table 5 The effect of using the BANNER entity tagger compared to gold-standard entities on the performance (precision, recall and F-score) of AkanePPI trained on AIMed.

From: A realistic assessment of methods for extracting gene/protein interactions from free text

 

(A)

B

H

I

L

Precision:

     

With gold-standard entities

57.0

29.2

61.5

60.2

69.6

With BANNER

34.3

23.8

26.8

13.0

39.8

Δ precision

22.7

5.4

34.7

47.2

29.8

Recall:

     

With gold-standard entities

74.0

31.8

44.2

32.5

23.8

With BANNER

64.2

30.2

41.1

24.5

27.4

Δ recall

9.8

1.7

3.1

8.0

-3.7

F-score:

     

With gold-standard entities

64.4

30.5

51.4

42.2

35.4

With BANNER

44.7

26.6

32.4

17.0

32.5

Δ F-score

19.7

3.8

19.0

25.2

3.0

  1. Corpus abbreviations are as follows: A = AIMed; B = BioInfer; H = HPRD50; I = IEPA; L = LLL.