Adaptive Fisher-Based Deep Convolutional Neural Network and Its Application to Recognition of Rolling Element Bearing Fault Patterns and Sizes

Table 3

The models adopted in this paper and the diagnosis results (No. 1 and No. 2 represent no. of testing samples and no. of correct identification, respectively).

Models	Model parameters	No. 1	No. 2	Calculation time (s)	Recognition rate (%)
DCNN	Prior knowledge	40	33	33.1260	82.5
FDCNN	Prior knowledge	40	37	63.1432	92.5
PSO-AFDCNN	Adaptively	40	40	123.4221	100
GSD-AFDCNN	Adaptively	40	40	96.2492	100