Table 2.
Both normal and abnormal signals of three diseases.
| Classification method | TPR of LQT1 | TPR of HCM | TPR of CPVT | Accuracy |
|---|---|---|---|---|
| kNN, cityblock metric, equal weighting, k = 1 | 86.7 | 87.3 | 80.7 | 83.2 |
| kNN, cityblock metric, inverse weighting, k = 5 | 94.4 | 84.5 | 79.0 | 83.5 |
| kNN, cityblock metric, squared inverse weighting, k = 5 | 91.1 | 85.9 | 81.5 | 84.5 |
| Random forests, 35 trees | 88.9 | 84.5 | 88.0 | 87.6 |
| LS-SVM cubic kernel, parameter C = 2−5 | 87.8 | 78.9 | 85.4 | 84.8 |
| LS-SVM RBF kernel, parameters C = 2, sigma = 2 | 90.0 | 78.9 | 88.4 | 87.1 |
True positive rates (TPR, %) for LQT1, HCM and CPVT diseases, with 90, 71 and 233 signals respectively, and accuracy (%) of all signals (kNN is k nearest neighbor searching method and LS-SVM RBF least square support vector machine with a radial basis function kernel). The best accuracies are bolded.