Skip to main content

Table 5 GE11 event extraction with scikit-learn classifiers.

From: TEES 2.2: Biomedical Event Extraction for Diverse Corpora

Classifier

Parameters

Recall

Precision

F-score

BernoulliNB

alpha = 0.001,0.01,0.1,1,10,100,1000

53.41

14.93

23.34

Perceptron

default

38.82

61.73

47.67

SVC

C=[10-3, 106], probability=True

47.06

66.05

54.96

LinearSVC

C=[10-3, 106]

46.65

68.02

55.35

ExtraTrees

n_estimators = 10,50,100

27.97

78.58

41.25

RandomForest

n_estimators = 10,50,100,500

24.92

78.65

37.84

SVMmulticlass

C=[1,106]

54.98

52.89

53.92

  1. The SVMmulticlass used in all TEES BioNLP Shared Task results is shown for reference. Performance is evaluated for task 1 on the development set with the downloadable GE11 evaluator.