Bearing Fault Diagnosis Based on Collaborative Representation Using Projection Dictionary Pair

Table 5

Recognition rates obtained by different classification methods on real vibration signals from accelerated degradation of bearing.

Method	KNN	SVM	DT	BPNN	Ours
Feature dimension	300	300	300	300	300
Acc (%)	95.4	97.6	89.5	92.6	100.0