Deep Adaptive Adversarial Network-Based Method for Mechanical Fault Diagnosis under Different Working Conditions

Table 2

Diagnosis results in Experiment 2.

Method	D⟶E	E⟶D	E⟶F	F⟶E	D⟶F	F⟶D	Average (%)
SAE	55.28 ± 6.85%	51.32 ± 4.36%	59.47 ± 4.20%	61.11 ± 3.54%	40.25 ± 5.36%	36.47 ± 4.69%	50.65
TCA	81.66 ± 4.51%	78.52 ± 5.23%	75.21 ± 4.20%	70.34 ± 6.85%	66.47 ± 7.87%	64.50 ± 5.24%	72.78
DANN	88.45 ± 4.81%	85.14 ± 4.15%	83.41 ± 5.14%	82.97 ± 5.97%	76.45 ± 4.78%	72.34 ± 5.97%	81.46
MK-MMD	94.75 ± 1.12%	93.84 ± 0.88%	91.63 ± 1.83%	92.47 ± 2.04%	88.62 ± 2.13%	85.31 ± 0.56%	91.10
SFDA	94.40 ± 0.07%	92.21 ± 1.02%	90.02 ± 0.81%	91.32 ± 1.24%	87.20 ± 0.34%	83.07 ± 0.61%	89.70
DAAN	95.41 ± 0.96%	94.96 ± 0.89%	93.81 ± 1.21%	94.07 ± 1.78%	89.74 ± 3.14%	87.96 ± 3.67%	92.65