def Predictor.fe_clean(self, data, subset=None) [source]
Validate and clean input data against trained standards.
Applies the trained data-quality pipeline: type
validation, outlier detection, and relevance filtering.
Variables that do not match the trained behavioral
profile are flagged (and optionally removed).
Parameters
data : pDF
Input data of shape (n_samples, n_features).
subset : list
Specific features to process and return.
Default = None: all trained features.
Default = None: all trained features.
Returns
FeCleanerResult