Principal Components of Superhigh-Dimensional Statistical Features and Support Vector Machine for Improving Identification Accuracies of Different Gear Crack Levels under Different Working Conditions

Table 2

The design of the three experiments (N, H, and F denotes no load, half load, and full load, resp.).

Experiment	Number of training/testing samples	Gear crack levels of training/testing	Motor speeds of training/testing samples (rpm)	Loads of training/testing samples	Classification labels
1	12/12	0%/0%	1200–1800/1200–1800	N, H, F/N, H, F	1
	12/12	25%/25%	1200–1800/1200–1800	N, H, F/N, H, F	2
	12/12	50%/50%	1200–1800/1200–1800	N, H, F/N, H, F	3
	12/12	75%/75%	1200–1800/1200–1800	N, H, F/N, H, F	4
	12/12	100%/100%	1200–1800/1200–1800	N, H, F/N, H, F	5
2	12/12	0%/0%	1200, 1600/1400, 1800	N, H, F/N, H, F	1
	12/12	25%/25%	1200, 1600/1400, 1800	N, H, F/N, H, F	2
	12/12	50%/50%	1200, 1600/1400, 1800	N, H, F/N, H, F	3
	12/12	75%/75%	1200, 1600/1400, 1800	N, H, F/N, H, F	4
	12/12	100%/100%	1200, 1600/1400, 1800	N, H, F/N, H, F	5
3	8/16	0%/0%	1200–1800/1200–1800	N/H, F	1
	8/16	25%/25%	1200–1800/1200–1800	N/H, F	2
	8/16	50%/50%	1200–1800/1200–1800	N/H, F	3
	8/16	75%/75%	1200–1800/1200–1800	N/H, F	4
	8/16	100%/100%	1200–1800/1200–1800	N/H, F	5