Imbalanced Fault Classification of Bearing via Wasserstein Generative Adversarial Networks with Gradient Penalty

Table 1

Description of the unbalanced Datasets A and B.

Healthcondition	The percent oftraining samples		The percent of testing samples
	Dataset A	Dataset B	Dataset A/B
NC	250	250	250
IF1	200	100	250
IF2	150	50	250
IF3	100	25	250
OF1	200	100	250
OF2	150	50	250
OF3	100	25	250
RF1	200	100	250
RF2	150	50	250
RF3	100	25	250