Skip to main content

Table 6 Results of the validation phase with probability score equal to 1.0. Performance prediction statistics are reported for the two subsets (“similar” and “not similar”) extracted from the validation dataset

From: Prediction of activity and selectivity profiles of human Carbonic Anhydrase inhibitors using machine learning classification models

 

TP

FN

FP

TN

Accuracy

MCC

Precision

Recall

hCA II, N = 350

Not similar

112

17

6

58

0.88

0.75

0.87

0.95

Similar

15

1

3

12

0.87

0.75

0.94

0.83

hCA IX, N = 450

Not similar

26

1

1

12

0.95

0.89

0.96

0.96

Similar

17

5

6

25

0.79

0.58

0.77

0.74

hCA XII, N = 150

Not similar

19

3

0

6

0.89

0.76

0.86

1.00

Similar

45

3

1

2

0.92

0.48

0.94

0.98

  1. Similarity with respect to training dataset was calculated to select similar (“similar”) or dissimilar (“not similar”) subsets. Performance metrics are reported in italics