Research Article
Optimal Statistical Feature Subset Selection for Bearing Fault Detection and Severity Estimation
Table 1
List of statistical time-domain features.
| Feature | Notation and formula |
| Mean T1 = | Root mean square T2 = | Root T3 = | Maximum value T4 = | Peak-to-peak T5 = | Standard deviation T6 = | Median T7 = | 25th percentiles T8 = | 75th percentiles T9 = | Skewness T10 = |
| Kurtosis T11 = | Crest factor T12 = | Shape factor T13 = | Impulse factor T14 = | Clearance factor T15 = | Skewness factor T16 = | Kurtosis factor T17 = | Geometric mean T18 = | Root-sum-of-squares (RSSQ) T19 = | Mean absolute deviation T20 = |
| Median absolute deviation T21 = | Rate of zero crossing T22 = | Entropy T23 = | Histogram upper bound T24 = | Histogram lower bound T25 = | Activity T26 = | Mobility T27 = | Complexity T28 = | Weibull negative log likelihood T29 = | Normal negative log likelihood for single Gaussian T30 = |
|
|