Skip to main content

Table 1 NER evaluation.

From: Assessment of disease named entity recognition on a corpus of annotated sentences

  

OMIM

Annotated

TP

Precision

Recall

F-measure

 

D look-up

908

1063

584

54.94

64.32

59.26

Exact boundary

Statistical

908

1051

274

26.07

30.18

27.97

 

MetaMap

908

873

273

31.27

30.07

30.66

 

D look-up

908

1063

598

56.26

65.86

60.68

Left alignment

Statistical

908

1051

390

39.86

38.33

39.08

 

MetaMap

908

873

348

37.11

42.95

39.82

 

D look-up

908

1063

685

64.44

75.44

69.51

Right alignment

Statistical

908

1051

439

53.49

51.43

52.44

 

MetaMap

908

873

467

41.77

48.35

44.82

This table shows the result of NER. For each alingment and method, first we show how many diseases are annotated in the corpus (OMIM) then the total number of annotation done by the method and then how many of them agreed with the benchmark (TP).