A Condition Monitoring Method of Hydraulic Gear Pumps Based on Multilevel Mechanism-Data Fusion

Table 6

Results of cross-validation of SVC models.

Kernel	Channel	Hyperparameters	Accuracy	Precision	Recall	F1-score	Successful proportion
SVC1 linear kernel			0.8519	0.9	0.75	0.8182	25/27
			0.8889	0.8462	0.9167	0.8800
			0.8889	0.8462	0.9167	0.8800
	Voted	/	0.9259	0.9167	0.9167	0.9167
SVC2 polynomial kernel			0.7407	0.7273	0.6667	0.6957	22/27
			0.8889	0.8462	0.9167	0.8800
			0.8519	0.7857	0.9167	0.8462
	Voted	/	0.8148	0.7692	0.8333	0.8
SVC3 Gaussian kernel			0.8519	1.0	0.6667	0.8	23/27
			0.8889	0.8462	0.9167	0.8800
			0.8519	0.8333	0.8333	0.8333
	Voted	/	0.8519	0.8333	0.8333	0.8333
SVC4 sigmoid kernel			0.8519	0.8333	0.8333	0.8333	23/27
			0.9259	1.0	0.8333	0.9091
			0.8519	0.8333	0.8333	0.8333
	Voted	/	0.8519	0.8333	0.8333	0.8333
SVC5 Laplacian kernel			0.8148	0.8889	0.6667	0.7619	23/27
			0.8519	0.8333	0.8333	0.8333
			0.8889	0.8462	0.9167	0.8800
	Voted	/	0.8519	0.8333	0.8333	0.8333