Table 23.
Dataset | vGRF Signal | CNN Classifier |
Time Window | Frequency Range | Classification Task | Evaluation Parameters | ||||
---|---|---|---|---|---|---|---|---|---|---|
Sen (%) | Spec (%) | Acc (%) | AUC | |||||||
All | LF | ResNet-101 | 10 s | 0.83–1.95 Hz | Multi-Class | Class 0 | 94.27 * | 97.68 * | 96.63 * | 0.9597 * |
ResNet-101 | 10 s | 0.83–1.95 Hz | Class 2 | 87.63 * | 94.06 * | 91.72 * | 0.9085 * | |||
ResNet-50 | 10 s | 0.83–1.95 Hz | Class 2.5 | 91.45 * | 93.35 * | 92.89 * | 0.9240 * | |||
GoogLeNet | 10 s | 1.95–50 Hz | Class 3 | 93.14 | 98.17 | 97.74 | 0.9566 | |||
ResNet-101 | 10 s | 0.83–1.95 Hz | Two-Class | 95.28 | 96.52 | 96.07 | 0.9915 | |||
RF | ResNet-50 | 10 s | 0.83–1.95 Hz | Multi-Class | Class 0 | 94.97 | 96.34 | 95.92 | 0.9566 | |
ResNet-101 | 10 s | 0.83–1.95 Hz | Class 2 | 87.88 | 93.72 | 91.59 | 0.9080 | |||
ResNet-50 | 15 s | 0.83–1.95 Hz | Class 2.5 | 89.53 | 94.82 | 93.55 | 0.9218 | |||
AlexNet | 10 s | 1.95–50 Hz | Class 3 | 94.95 * | 97.30 * | 97.09 * | 0.9612 * | |||
ResNet-101 | 10 s | 0.83–1.95 Hz | Two-Class | 94.46 * | 97.69 * | 96.63 * | 0.9949 * | |||
CF | ResNet-50 | 10 s | 0.83–1.95 Hz | Multi-Class | Class 0 | 91.65 | 96.21 | 94.81 | 0.9393 | |
ResNet-50 | 10 s | 0.83–1.95 Hz | Class 2 | 87.37 | 91.19 | 89.80 | 0.8928 | |||
ResNet-50 | 10 s | 0.83–1.95 Hz | Class 2.5 | 85.20 | 95.55 | 93.04 | 0.9038 | |||
AlexNet | 10 s | 1.95–50 Hz | Class 3 | 92.06 | 98.61 | 98.05 | 0.9534 | |||
ResNet-101 | 10 s | 0.83–1.95 Hz | Two-Class | 92.71 | 95.11 | 94.37 | 0.9856 |
Note: * denotes the best classification result and was selected using Youden’s index criteria.