Research Article
Discriminant Feature Distribution Analysis-Based Hybrid Feature Selection for Online Bearing Fault Diagnosis in Induction Motors
Table 7
Classification performance of the three different approaches for standard benchmark data with white Gaussian noise of 15 dB.
| Model | 14 mil 1730 rpm | 14 mil 1750 rpm | 14 mil 1772 rpm | 14 mil 1797 rpm | 21 mil 1730 rpm | 21 mil 1750 rpm | 21 mil 1772 rpm | 21 mil 1797 rpm |
| All features | 83.89 | 90.42 | 90.42 | 83.89 | 100.00 | 96.81 | 92.67 | 98.33 | Proposed model | 91.81 | 93.19 | 93.89 | 91.19 | 100.00 | 99.58 | 96.69 | 99.86 | Algorithm [25] | 84.20 | 83.19 | 90.83 | 82.30 | 100.00 | 96.67 | 93.10 | 98.89 |
|
|