Table 2.

Comparison of classification accuracy on simulated database (using traditional methods and deep-learning-based multilayer perceptron (MLP) model). PCA, principal component analysis; FCM, fuzzy C-means; FSDE, first and second derivative extrema; CORR, correlation-based; SVM, support vector machine.

Dataset	NL 1	PCA + FCM		FSDE + K-means	CORR + FCM		Fusion + SVM	MLP	1D-CNN
		F 2 = 3	F = 10	F = 3	F = 3	F = 10	F = 10		E2 3
C_Easy1	005	99.37	99.35	94.62	97.50	97.38	98.66	99.26	99.53
	010	99.72	99.72	95.54	94.04	96.45	98.98	99.43	99.72
	015	99.25	99.28	94.45	90.54	94.94	98.22	99.25	99.49
	020	99.40	99.40	95.08	88.77	92.43	97.35	99.19	99.33
	025	99.24	99.24		84.41	86.84	95.45		99.56
	030	98.73	98.59		81.50	80.83	88.66		99.65
	035	97.76	95.16		77.02	73.80	83.22		99.43
	040	96.49	68.54		75.58	64.62	78.12		99.51
C_Easy2	005	98.48	98.68	94.81	93.20	96.04	92.23	98.68	99.64
	010	97.16	98.24	94.83	86.02	82.19	92.93	98.49	99.62
	015	92.52	94.49	94.96	83.05	82.82	89.80	97.19	98.89
	020	85.20	88.60	92.71	79.81	78.22	86.24	95.20	99.69
C_Difficult1	005	95.86	72.54	94.50	83.48	86.08	97.58	98.78	99.21
	010	89.56	66.11	94.78	65.69	71.55	94.81	98.93	99.00
	015	76.41	61.33	93.81	57.49	58.84	87.85	97.55	98.43
	020	63.03	54.05	90.60	53.72	53.81	78.59	96.62	98.54
C_Difficult2	005	98.69	98.81	94.38	91.50	94.50	87.40	98.49	98.88
	010	98.64	98.76	94.48	90.96	96.33	88.07	94.66	99.78
	015	94.39	97.33	87.18	88.17	96.02	74.65	82.20	99.58
	020	84.63	83.37	81.71	84.77	95.48	67.25	51.55	99.75

¹ Noise level; ² feature dimensions; ³ experiment.