AbstractNumericOutlierDetector 
Ancestor for numeric outlier detection.

AbstractOutlierDetector 
Ancestor for outlier detectors of actual vs predicted.

IQR 
Interquartile range based detector.
If difference between actual/predicted is more than the factor of standard deviations away from the quartial 0.25/0.75, then the point gets flagged as outlier.

MeanAbsoluteError 
Mean absolute error (MAE) based detector.
If difference between actual/predicted is more than MAE * FACTOR, then the point gets flagged as outlier.

MultiOutlierDetector 
Applies the specified outlier detectors sequentially and combines the detected outliers either via union or intersect.

Null 
Dummy outlier detector, detects no outliers at all.

StdDev 
Standard deviation based detector.
If difference between actual/predicted is more than the factor of standard deviations away from the mean, then the point gets flagged as outlier.
