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

Table 1

Classification results in Experiment 1.

Method	A ⟶ B	B ⟶ A	B ⟶ C	C ⟶ B	A ⟶ C	C ⟶ A	Average
SAE	73.25 ± 1.48%	71.19 ± 2.01%	74.12 ± 3.17%	62.64 ± 3.51%	46.45 ± 4.97%	43.53 ± 4.35%	60.20
TCA	92.05 ± 2.94%	88.24 ± 3.13%	85.19 ± 2.47%	83.57 ± 1.94%	72.51 ± 5.35%	67.63 ± 6.20%	81.53
DANN	95.24 ± 1.54%	93.45 ± 1.71%	89.34 ± 2.84%	87.84 ± 2.64%	79.78 ± 4.40%	75.64 ± 4.21%	86.88
MK-MMD	98.04 ± 0.12%	96.87 ± 0.88%	95.84 ± 1.54%	96.14 ± 1.75%	91.97 ± 1.58%	90.54 ± 1.96%	94.90
SFDA	97.85 ± 0.25%	95.87 ± 1.04%	92.82 ± 0.85%	94.35% ± 1.09%	89.45 ± 0.83%	86.55 ± 2.04%	92.82
DAAN	98.63 ± 0.63%	97.25 ± 0.90%	97.02 ± 1.68%	96.41 ± 1.86%	92.34 ± 2.65%	90.98 ± 3.21%	95.44