Table 5.
SVM-based prediction performances for four different types of prediction methods using equal positive and negative instances
|
Prediction |
Binary approach |
PSSM approach |
||||||
|---|---|---|---|---|---|---|---|---|
| Sensitivity | Specificity | Accuracy | MCC | Sensitivity | Specificity | Accuracy | MCC | |
| VIRs |
65.98 ± 0.85 |
65.85 ± 0.52 |
65.91 ± 0.60 |
0.32 ± 0.01 |
75.80 ± 0.35 |
77.07 ± 0.69 |
76.43 ± 0.47 |
0.53 ± 0.01 |
| VAIRs |
62.09 ± 2.01 |
61.87 ± 2.92 |
61.99 ± 1.30 |
0.24 ± 0.03 |
73.25 ± 2.43 |
73.83 ± 0.95 |
73.54 ± 1.47 |
0.47 ± 0.03 |
| VBIRs |
68.55 ± 0.75 |
68.37 ± 0.83 |
68.47 ± 0.44 |
0.37 ± 0.01 |
80.08 ± 0.61 |
82.49 ± 0.79 |
81.29 ± 0.23 |
0.63 ± 0.01 |
| PLPIRs | 76.74 ± 1.73 | 74.91 ± 1.42 | 75.82 ± 1.32 | 0.52 ± 0.03 | 89.85 ± 0.87 | 89.85 ± 1.16 | 89.84 ± 0.70 | 0.80 ± 0.01 |
The values of standard errors are also given with performances.