Table 22.

Multi-class (CO (Class 0) vs. PD Stage 2 (Class 2) vs. PD Stage 2.5 (Class 2.5), and PD Stage 3 (Class 3)) and two-class classification (CO vs. PD (Class 2, Class 2.5, and Class 3)) summary using several CNN classifiers (AlexNet, ResNet-50, ResNet-101, and GoogLeNet) with 10-fold cross-validation for Si dataset.

Dataset	vGRF Signal	CNN Classifier	Time Window	Frequency Range	Classification Task		Evaluation Parameters
							Sen (%)	Spec (%)	Acc (%)	AUC
Si	LF	AlexNet	10 s	1.95–50 Hz	Multi-Class	Class 0	98.56 *	99.05 *	98.83 *	0.9881 *
		AlexNet	10 s	1.95–50 Hz		Class 2	95.83 *	98.84 *	97.53 *	0.9734 *
		AlexNet	10 s	1.95–50 Hz		Class 2.5	98.81 *	98.39 *	98.44 *	0.9860 *
		ResNet-50	10 s	0.83–1.95 Hz	Two-Class		98.85 *	98.41 *	98.56 *	0.9964 *
	RF	ResNet-101	10 s	0.83–1.95 Hz	Multi-Class	Class 0	99.43	97.86	98.57	0.9864
		ResNet-101	10 s	0.83–1.95 Hz		Class 2	95.83	97.22	96.61	0.9653
		ResNet-50	15 s	1.95–50 Hz		Class 2.5	94.64	98.03	97.66	0.9633
		ResNet-50	10 s	0.83–1.95 Hz	Two-Class		98.62	98.63	98.57	0.9994
	CF	AlexNet	10 s	1.95–50 Hz	Multi-Class	Class 0	97.41	99.52	98.57	0.9847
		AlexNet	10 s	1.95–50 Hz		Class 2	94.94	97.69	96.48	0.9631
		AlexNet	15 s	1.95–50 Hz		Class 2.5	98.21	97.37	97.46	0.9779
		ResNet-101	10 s	0.83–1.95 Hz	Two-Class		98.88	97.35	97.92	0.9984

Note: * denotes the best classification result and was selected using Youden’s index criteria.