Skip to main content

Table 3 Performance of classification model under different evaluation strategies.

From: Prediction of neddylation sites from protein sequences and sequence-derived properties

Evaluation Strategy

Acc

Sp

Sn

MCC

AUC

Self-consistency

0.92

0.92

0.94

0.56

0.98

5-fold stratified cross-validation

0.91

0.91

0.75

0.44

0.95

10-fold stratified cross-validation

0.91

0.91

0.76

0.45

0.95

Validation set

0.90

0.91

0.67

0.39

0.83

Holdout set

0.90

0.91

0.64

0.35

0.80

  1. Cross-validation (CV) results were reported as means of 100 repeats. As two standard errors were not exceeding 0.01, they were not reported.