A Review of Artificial Intelligence Methods for Condition Monitoring and Fault Diagnosis of Rolling Element Bearings for Induction Motor

Table 4

DL vs. ML for bearing CM and FDD.

The point	DL algorithms	Conventional ML algorithms
Feature extraction	Automated	Manual
Classification	Simultaneous feature learning and classification	Feature extraction and classification are separated
Human expertise	Not required	Required
Prior knowledge	Not required	Required
External noise	Better	Sensitive
Frequent change	Better	Sensitive
Diagnostic accuracy	High accuracy	Lesser accuracy
Dataset	Efficient for big data	Efficient for small datasets
Computational training cost	Expensive	Lesser computational training cost is required
Hardware requirements	GPU is required	CPU is efficient