Skip to main content

Table 8 Performance metrics values and corresponding confusion tables for three different blending methods

From: Comparison and improvement of the predictability and interpretability with ensemble learning models in QSPR applications

 

Training set

Test set

Acc (%)

F1 (%)

Acc (%)

Pr (%)

r (%)

F1 (%)

MCC (%)

Actual class

Predicted class

LC

NLC

Uniform blending

99.5

99.7

88.3

91.3

93.6

92.5

67.3

LC

674

46

NLC

64

163

Linear blending

99.5

99.7

88.4

91.6

93.5

92.5

67.8

LC

673

47

NLC

62

165

Any blending

99.3

99.5

88.8

91.7

93.8

92.7

68.6

LC

675

45

NLC

61

166