Table 4. Best results of SVM on the selected feature subsets.
| No. of feature | Feature subset | SVM | ||
|---|---|---|---|---|
| Average accuracy (%) | Root Mean Square Error (RMSE) | Average AUROC | ||
| 10 | Eth Smo Dri Chew Diff Inv Node Tre p53 p63 | 64.76 | 0.4000 | 0.5000 |
| 9 | Eth Smo Dri Chew Diff Inv Tre p53 p63 | 64.76 | 0.4000 | 0.5000 |
| 8 | Eth Smo Dri Chew Diff Inv p53 p63 | 64.76 | 0.4000 | 0.5000 |
| 7 | Eth Smo Dri Chew Diff Inv p63 | 64.76 | 0.4000 | 0.5000 |
| 6 | Smo Dri Chew Sub Diff p63 | 61.43 | 0.6000 | 0.4750 |
| 5 | Smo Dri Chew Diff p63 | 64.76 | 0.4000 | 0.5000 |
| 4 | Smo Chew Diff p63 | 64.76 | 0.4000 | 0.5000 |
| 3 | Smo Diff p63 | 64.76 | 0.4000 | 0.5000 |
| 2 | Smo p63 | 61.43 | 0.4000 | 0.5000 |