cat2cat.cat2cat_ml_utils
========================

.. py:module:: cat2cat.cat2cat_ml_utils


Functions
---------

.. autoapisummary::

   cat2cat.cat2cat_ml_utils.safe_nanmean
   cat2cat.cat2cat_ml_utils.resolve_ml_models
   cat2cat.cat2cat_ml_utils.prepare_ml_frames
   cat2cat.cat2cat_ml_utils.brier_score
   cat2cat.cat2cat_ml_utils.mean_true_probability
   cat2cat.cat2cat_ml_utils.apply_ml_fallback
   cat2cat.cat2cat_ml_utils._feature_is_direct
   cat2cat.cat2cat_ml_utils._weight_column_method_name


Module Contents
---------------

.. py:function:: safe_nanmean(values: Sequence[Any], ndigits: int = 3) -> float

.. py:function:: resolve_ml_models(ml: cat2cat.dataclass.cat2cat_ml) -> List[Tuple[str, sklearn.base.ClassifierMixin]]

.. py:function:: prepare_ml_frames(ml: cat2cat.dataclass.cat2cat_ml, target_data: Optional[pandas.DataFrame] = None) -> Tuple[pandas.DataFrame, Optional[pandas.DataFrame], List[str]]

.. py:function:: brier_score(prob_matrix: pandas.DataFrame, true_cats: Sequence[Any], classes: Sequence[Any]) -> float

.. py:function:: mean_true_probability(prob_matrix: pandas.DataFrame, true_cats: Sequence[Any]) -> float

.. py:function:: apply_ml_fallback(target_df: pandas.DataFrame, ml_names: Sequence[str], on_fail: str, fail_warn: bool) -> None

.. py:function:: _feature_is_direct(feature: Any) -> bool

.. py:function:: _weight_column_method_name(ml_name: str) -> str