Skip to main content

Table 1 Performance comparisons between base learner and ensemble models on D92M over leave-one-out cross-validation. All the results are obtained by setting 0.5 as the default discrimination threshold to generate the prediction label except the Maximum MCC values which are obtained by identifying the thresholds that maximize the values of MCC

From: Identifying GPCR-drug interaction based on wordbook learning from sequences

Metrics

Base learner

Ensemble model with different Ne

2

4

6

8

10

Acc (%)

83.60

84.09

85.05

84.25

84.41

84.73

Sn (%)

81.42

80.31

81.1

80.47

80.63

80.94

Sp (%)

84.73

86.04

87.1

86.2

86.37

86.69

MCC

0.64

0.65

0.67

0.66

0.66

0.66

Maximum MCC

0.65

0.68

0.68

0.69

0.69

0.69