Research Article
Iterative Reweighted Noninteger Norm Regularizing SVM for Gene Expression Data Classification
Table 4
Comparison of the prediction error (Pe%) and feature selection error (Fse%) on all four clinical cancer datasets.
| Datasets | IRWP-SVM | L2-SVM | L1-SVM | L0-SVM | Random forest |
| Bladder | | | | | | Pe | 20.1 | 23.6 | 26.41 | 28.41 | 22.56 | Fse | 3.47 | 33 | 4.21 | 9.28 | — | Melanoma | | | | | | Pe | 13.68 | 14.71 | 14.11 | 18.13 | 11.13 | Fse | 0.25 | 23.8 | 1.5 | 6.34 | — | Lymphoma | | | | | | Pe | 6.67 | 7.39 | 6.10 | 10.10 | 3.28 | Fse | 1. 1 | 14.3 | 1.62 | 5.91 | — | Colon | | | | | | Pe | 10 | 12.2 | 16.39 | 11.18 | 19.75 | Fse | 2.34 | 14.7 | 3.10 | 8.21 | — |
|
|