A Novel Method for Diagnosis of Bearing Fault Using Hierarchical Multitasks Convolutional Neural Networks

Table 4

Performance comparison of accuracy metric between other models and the proposed HMCNN model on bearing fault dataset.

Case	Number training samples	Number testing samples	Accuracy
			SVM	BPNN	CNN	LSTM	HMCNN
1	500	500	0.716	0.682	0.856	0.846	0.967
2	1000	500	0.753	0.712	0.897	0.886	0.978
3	1500	500	0.761	0.842	0.912	0.893	0.991
4	2000	500	0.762	0.856	0.921	0.906	0.997