Missforest wrapper

fyt.wrappers.missforest_wrapper

MissForestWrapped

Bases: BaseEstimator, TransformerMixin

A wrapper around MissForest for polars/pandas DataFrame I/O and sklearn compatibility.

This wrapper extends the MissForest imputer from the missforest library to be compatible with scikit-learn pipelines. Accepts polars DataFrames as input, converts internally to pandas for MissForest, and returns polars DataFrames as output.

Parameters:

Name	Type	Description	Default
clf	BaseEstimator | None	Classifier estimator to use for categorical features (optional). If None, uses the default MissForest classifier (LightGBM).	None
rgr	BaseEstimator | None	Regressor estimator to use for numerical features (optional). If None, uses the default MissForest regressor (LightGBM).	None
categorical	list[str] | None	List of categorical column names (optional). These columns will be identified and handled appropriately during imputation.	None
random_state	int | None	Seed for reproducible imputation (optional). MissForest itself has no seed parameter; randomness enters through its internal LightGBM estimators. When random_state is set and no custom estimators are provided, seeded LGBMClassifier/LGBMRegressor instances are built and passed to MissForest so that repeated runs produce identical imputations. Ignored when both clf and rgr are provided (seed those estimators directly instead).	None
**kwargs	object	Additional parameters to pass to the MissForest constructor (e.g., max_iter, initial_guess, etc.).	{}

Example

Using custom estimators (created via dependency injection)

from sklearn.ensemble import RandomForestClassifier, RandomForestRegressor

clf = RandomForestClassifier(n_estimators=100, random_state=42) rgr = RandomForestRegressor(n_estimators=100, random_state=42)

imputer = MissForestWrapped( clf=clf, rgr=rgr, categorical=["sex", "smoker"], max_iter=10 )

Using default estimators

imputer = MissForestWrapped(categorical=["sex", "smoker"])

__getattr__(name)

Delegate attribute access to the underlying MissForest instance.

Parameters:

Name	Type	Description	Default
name	str	Attribute name.	required

Returns:

Type	Description
	Attribute value from MissForest instance.

Raises:

Type	Description
AttributeError	If attribute not found.

fit(X, y=None)

Fit the MissForest imputer.

Parameters:

Name	Type	Description	Default
X	DataFrame | DataFrame	polars or pandas DataFrame with missing values.	required
y	object	Ignored, present for sklearn compatibility.	None

Returns:

Name	Type	Description
self	MissForestWrapped	Fitted transformer.

fit_transform(X, y=None, **fit_params)

Fit the imputer and transform the data.

Parameters:

Name	Type	Description	Default
X	DataFrame | DataFrame	polars or pandas DataFrame with missing values.	required
y	object	Ignored, present for sklearn compatibility.	None
**fit_params	object	Additional fit parameters forwarded to fit().	{}

Returns:

Name	Type	Description
X_imputed		polars DataFrame with imputed values.

get_params(deep=True)

Get parameters for this estimator.

Parameters:

Name	Type	Description	Default
deep	bool	If True, will return the parameters for this estimator and contained subobjects that are estimators.	True

Returns:

Type	Description
dict[str, object]	Dictionary of parameter names mapped to their values.

set_output(*, transform=None)

Set output configuration for compatibility with sklearn pipelines.

Parameters:

Name	Type	Description	Default
transform	str | None	Output format for transform method. Not used in this wrapper.	None

Returns:

Name	Type	Description
MissForestWrapped	MissForestWrapped	The transformer itself.

set_params(**params)

Set the parameters of this estimator.

Parameters:

Name	Type	Description	Default
**params	object	Estimator parameters.	{}

Returns:

Name	Type	Description
self	MissForestWrapped	Estimator instance.

transform(X)

Transform the data using fitted MissForest parameters.

Parameters:

Name	Type	Description	Default
X	DataFrame | DataFrame	polars or pandas DataFrame with missing values.	required

Returns:

Name	Type	Description
X_imputed		polars DataFrame with imputed values.

Raises:

Type	Description
RuntimeError	If the imputer has not been fitted yet.