Back to models
Transformer

HampelFilter

API referenceSource
Missing valuesUnequal length

Use HampelFilter to detect outliers based on a sliding window.

Quickstart

python
from sktime.transformations.outlier_detection import HampelFilter

estimator = HampelFilter(window_length=10, n_sigma=3, k=1.4826, return_bool=False)

Parameters(4)

window_lengthint, optional (default=10)
Length of the sliding window
n_sigmaint, optional (default=3)
Defines how strong a point must outly to be an “outlier”
kfloat, optional (default = 1.4826)
A constant scale factor which is dependent on the distribution, for Gaussian it is approximately 1.4826, by default 1.4826
return_boolbool, optional (default=False)
If True, outliers are filled with True and non-outliers with False. Else, outliers are filled with np.nan.

Examples

>>> from sktime.transformations.outlier_detection import HampelFilter
>>> from sktime.datasets import load_airline
>>> y = load_airline ()
>>> transformer = HampelFilter (window_length = 10)
>>> y_hat = transformer. fit_transform (y)

References

  1. [1 ] Hampel F. R., “The influence curve and its role in robust estimation”, Journal of the American Statistical Association, 69, 382-393, 1974