Skip to main content

Table 4 Prediction performance of the neural network model with different units in the hidden layer via 10-fold stratified cross-validation test

From: SIMLIN: a bioinformatics tool for prediction of S-sulphenylation in the human proteome based on multi-stage ensemble-learning models

#Units in the hidden layer

Decay

AUC

Sensitivity

Specificity

1

0

0.999842 ± 3.15E-4

0.999685 ± 6.30E-4

1

0.0004

0.999994 ± 6.30E-5

0.999887 ± 3.62E-4

1

0.1

1

0.999874 ± 3.68E-4

1

3

0

0.999874 ± 3.35E-4

0.999723 ± 6.84E-4

1

0.0004

0.999987 ± 8.85E-5

0.999937 ± 2.76E-4

1

0.1

1

0.999874 ± 3.80E-4

1

5

0

0.999793 ± 5.90E-4

0.999685 ± 7.02E-4

0.999902 ± 9.80E-4

0.0004

0.999869 ± 7.28E-4

0.999912 ± 4.48E-4

0.999704 ± 2.20E-3

0.1

1

0.999899 ± 3.44E-4

1