Table 3 Prediction performance of different classifiers for vitamin B-interacting residues (VBIRs)
Feature |     Classifier | SN | SP | ACC | MCC |
|---|---|---|---|---|---|
 Binary | SVM (Threshold = −0.8) | 73.22 ± 0.36 | 67.00 ± 0.49 | 67.57 ± 0.47 | 0.24 ± 0.00 |
SVM (Threshold = −0.6) | 30.36 ± 0.62 | 96.69 ± 0.12 | 90.66 ± 0.11 | 0.33 ± 0.01 | |
BayesNet | 63.25 ± 0.56 | 66.23 ± 0.73 | 65.96 ± 0.62 | 0.18 ± 0.00 | |
ComplementNaiveBayes | 68.69 ± 0.52 | 68.51 ± 0.23 | 68.52 ± 0.18 | 0.23 ± 0.00 | |
NaiveBayes | 37.74 ± 0.90 | 90.45 ± 0.23 | 85.66 ± 0.14 | 0.25 ± 0.01 | |
NaiveBayesMultinomial | 44.22 ± 0.43 | 87.54 ± 0.24 | 83.60 ± 0.19 | 0.25 ± 0.00 | |
IBk | 30.81 ± 0.71 | 93.33 ± 0.17 | 87.65 ± 0.14 | 0.24 ± 0.01 | |
|  | RandomForest | 39.33 ± 1.08 | 79.36 ± 0.37 | 75.72 ± 0.36 | 0.13 ± 0.01 |
 PSSM | SVM (Threshold = −0.8) | 83.33 ± 0.36 | 80.51 ± 0.13 | 80.77 ± 0.14 | 0.42 ± 0.00 |
SVM (Threshold =0.1) | 55.57 ± 0.63 | 98.04 ± 0.10 | 94.18 ± 0.09 | 0.61 ± 0.01 | |
BayesNet | 71.65 ± 1.13 | 66.14 ± 0.08 | 66.64 ± 0.10 | 0.23 ± 0.01 | |
ComplementNaiveBayes | 63.90 ± 1.26 | 81.73 ± 0.28 | 80.11 ± 0.22 | 0.32 ± 0.01 | |
NaiveBayes | 72.28 ± 1.22 | 66.44 ± 0.09 | 66.97 ± 0.12 | 0.23 ± 0.01 | |
NaiveBayesMultinomial | 21.22 ± 0.69 | 98.88 ± 0.03 | 91.82 ± 0.06 | 0.34 ± 0.01 | |
IBk | 56.74 ± 0.80 | 98.04 ± 0.07 | 94.28 ± 0.11 | 0.62 ± 0.01 | |
|  | RandomForest | 39.16 ± 0.56 | 97.74 ± 0.09 | 92.41 ± 0.10 | 0.46 ± 0.01 |