Table 2.
Classifiers Trained on Feature Sets Selected by Statistical Approaches
| Method | No. of Features | CV Accuracy | CV p value |
|---|---|---|---|
| SP | 1,124 | 0.98 | 0.91 |
| HE | 162 | 0.98 | 0.91 |
| SNET | 21 | 0.84 | 0.06 |
| FSNET | 36 | 0.96 | 0.06 |
| PFSNET | 65 | 0.92 | 0.06 |
All approaches have high CV accuracy, but this does not mean that this good CV accuracy in itself is meaningful or specific.