Table 6. MLP, RBF neural networks and linear SVM (), SVM with polynomial kernel () and SVM with RBF kernel () classifications using the projections of the features from PCA analysis of the HRV records from normal (N) and cardiovascular risk (R) subjects.
Number of PCA Projections | Classifier | Se (%) | Sp (%) | Np (%) | Pp (%) | Ac (%) |
3 | MLP | 60.00 | 73.33 | 64.71 | 69.23 | 66.67 |
RBFNN | 40.00 | 93.33 | 60.87 | 85.71 | 66.67 | |
SVM (Linear) | 77.27 | 63.64 | 73.68 | 68.00 | 70.45 | |
SVM (Polynomial kernel) | 84.09 | 70.45 | 81.58 | 74.00 | 77.27 | |
SVM (RBF kernel) | 89.39 | 65.15 | 86.00 | 71.95 | 77.27 | |
4 | MLP | 80.00 | 80.00 | 80.00 | 80.00 | 80.00 |
RBFNN | 46.67 | 86.67 | 61.90 | 77.78 | 66.67 | |
SVM (Linear) | 90.91 | 90.91 | 90.91 | 90.91 | 90.91 | |
SVM (Polynomial kernel) | 90.91 | 93.18 | 91.11 | 93.02 | 92.05 | |
SVM (RBF kernel) | 93.94 | 89.39 | 93.65 | 89.86 | 91.67 | |
5 | MLP | 66.67 | 100.00 | 75.00 | 100.00 | 83.33 |
RBFNN* | N/A | N/A | N/A | N/A | N/A | |
SVM (Linear) | 86.36 | 95.45 | 87.50 | 95.00 | 90.91 | |
SVM (Polynomial kernel) | 81.82 | 95.45 | 84.00 | 94.74 | 88.64 | |
SVM (RBF kernel) | 81.82 | 95.45 | 84.00 | 94.74 | 88.64 | |
10 | MLP | 80.00 | 73.33 | 78.57 | 75.00 | 76.67 |
RBFNN* | N/A | N/A | N/A | N/A | N/A | |
SVM (Linear) | 81.82 | 81.82 | 81.82 | 81.82 | 81.82 | |
SVM (Polynomial kernel) | 81.82 | 95.45 | 84.00 | 94.74 | 88.64 | |
SVM (RBF kernel) | 81.82 | 95.45 | 84.00 | 94.74 | 88.64 |
*The results for this classifier were not reported due to their poorness.