captest.capdata.CapData.filter_outliers

CapData.filter_outliers(custom_name=None, **kwargs)

Apply EllipticEnvelope from scikit-learn to remove outliers in (poa, power).

Parameters:

  • custom_name (str, default None) – Optional display label for the recorded filter step.

  • **kwargs – Forwarded to sklearn.covariance.EllipticEnvelope. Defaults support_fraction=0.9 and contamination=0.04 are applied when not overridden.

Notes

When NaN values are present in poa/power, filter_missing is invoked first (and recorded as a separate filter step). This preserves the legacy auto-handling behavior.