Table 2. Classification performance of different classifiers* on a training and test set using 10-fold cross-validation.
| LDA | QDA | RDA | MDA | FDA | NB | SVM | BAG | BOO | RF | |
|---|---|---|---|---|---|---|---|---|---|---|
| TRAIN MCR | 10.6 | 9.3 | 9.7 | 7.9 | 9.25 | 10.6 | 7.9 | 7.8 | 9.3 | 6.0 |
| TRAIN FNR | 7.9 | 0.9 | 7.0 | 4.1 | 4.2 | 5.1 | 0.9 | 2.3 | 4.6 | 1.9 |
| TEST MCR | 13.9 | 11.1 | 15.7 | 11.1 | 10.2 | 11.1 | 4.6 | 12.0 | 13.9 | 10.2 |
| TEST FNR | 10.2 | 4.6 | 13.0 | 7.4 | 5.6 | 6.5 | 0.9 | 5.6 | 7.4 | 4.6 |
*Settings: RDA: lambda 0.8997867, MDA: 3 subclasses, NB: Epanechnikov, SVM: Gaussian, BAG: 50 trees, BOO: 4561 iterations, RF: 500 trees, 3 features)