Reliable Fault Diagnosis of Rotary Machine Bearings Using a Stacked Sparse Autoencoder-Based Deep Neural Network

Table 5

The average classification accuracy of the proposed method, BPNN, VSI, and SDA for variable shaft speed conditions.

Method	Training data RPM	Testing data RPM	Average classification accuracy (%)
VSI	1796	1772, 1748, 1722	63.33
	1772	1796, 1748, 1722	85
	1748	1796, 1772, 1722	60
	1722	1796, 1772, 1748	67
ANN	1796	1772, 1748, 1722	83
	1772	1796, 1748, 1722	58
	1748	1796, 1772, 1722	87
	1722	1796, 1772, 1748	53
SDA	1796	1772, 1748, 1722	65
	1772	1796, 1748, 1722	49
	1748	1796, 1772, 1722	46.66
	1722	1796, 1772, 1748	46
Proposed	1796	1772, 1748, 1722	96.33
	1772	1796, 1748, 1722	93
	1748	1796, 1772, 1722	96
	1722	1796, 1772, 1748	90