OrdinalEncoder #

classsklearn.preprocessing.OrdinalEncoder(*,categories='auto',dtype=<class'numpy.float64'>,handle_unknown='error',unknown_value=None,encoded_missing_value=nan,min_frequency=None,max_categories=None)[source]#

Encode categorical features as an integer array.

The input to this transformer should be an array-like of integers orstrings, denoting the values taken on by categorical (discrete) features.The features are converted to ordinal integers. This results ina single column of integers (0 to n_categories - 1) per feature.

Read more in theUser Guide.For a comparison of different encoders, refer to:Comparing Target Encoder with Other Encoders.

Added in version 0.20.

Parameters:

categories‘auto’ or a list of array-like, default=’auto’

Categories (unique values) per feature:

‘auto’ : Determine categories automatically from the training data.
list :categories[i] holds the categories expected in the ithcolumn. The passed categories should not mix strings and numericvalues, and should be sorted in case of numeric values.

The used categories can be found in thecategories_ attribute.

dtypenumber type, default=np.float64

Desired dtype of output.

handle_unknown{‘error’, ‘use_encoded_value’}, default=’error’

When set to ‘error’ an error will be raised in case an unknowncategorical feature is present during transform. When set to‘use_encoded_value’, the encoded value of unknown categories will beset to the value given for the parameterunknown_value. Ininverse_transform, an unknown category will be denoted as None.

Added in version 0.24.

unknown_valueint or np.nan, default=None

When the parameter handle_unknown is set to ‘use_encoded_value’, thisparameter is required and will set the encoded value of unknowncategories. It has to be distinct from the values used to encode any ofthe categories infit. If set to np.nan, thedtype parameter mustbe a float dtype.

Added in version 0.24.

encoded_missing_valueint or np.nan, default=np.nan

Encoded value of missing categories. If set tonp.nan, then thedtypeparameter must be a float dtype.

Added in version 1.1.

min_frequencyint or float, default=None

Specifies the minimum frequency below which a category will beconsidered infrequent.

Ifint, categories with a smaller cardinality will be consideredinfrequent.
Iffloat, categories with a smaller cardinality thanmin_frequency*n_samples will be considered infrequent.

Added in version 1.3:Read more in theUser Guide.

max_categoriesint, default=None

Specifies an upper limit to the number of output categories for each inputfeature when considering infrequent categories. If there are infrequentcategories,max_categories includes the category representing theinfrequent categories along with the frequent categories. IfNone,there is no limit to the number of output features.

max_categories donot take into account missing or unknowncategories. Settingunknown_value orencoded_missing_value to aninteger will increase the number of unique integer codes by one each.This can result in up tomax_categories+2 integer codes.

Added in version 1.3:Read more in theUser Guide.

Attributes:

categories_list of arrays: The categories of each feature determined duringfit (in order ofthe features in X and corresponding with the output oftransform).This does not include categories that weren’t seen duringfit.
n_features_in_int: Number of features seen duringfit.
Added in version 1.0.
feature_names_in_ndarray of shape (n_features_in_,): Names of features seen duringfit. Defined only whenXhas feature names that are all strings.
Added in version 1.0.
infrequent_categories_list of ndarray: Infrequent categories for each feature.

See also

OneHotEncoder: Performs a one-hot encoding of categorical features. This encoding is suitable for low to medium cardinality categorical variables, both in supervised and unsupervised settings.
TargetEncoder: Encodes categorical features using supervised signal in a classification or regression pipeline. This encoding is typically suitable for high cardinality categorical variables.
LabelEncoder: Encodes target labels with values between 0 andn_classes-1.

Notes

With a high proportion ofnan values, inferring categories becomes slow withPython versions before 3.10. The handling ofnan values was improvedfrom Python 3.10 onwards, (c.f.bpo-43475).

Examples

Given a dataset with two features, we let the encoder find the uniquevalues per feature and transform the data to an ordinal encoding.

>>>fromsklearn.preprocessingimportOrdinalEncoder>>>enc=OrdinalEncoder()>>>X=[['Male',1],['Female',3],['Female',2]]>>>enc.fit(X)OrdinalEncoder()>>>enc.categories_[array(['Female', 'Male'], dtype=object), array([1, 2, 3], dtype=object)]>>>enc.transform([['Female',3],['Male',1]])array([[0., 2.],       [1., 0.]])

>>>enc.inverse_transform([[1,0],[0,1]])array([['Male', 1],       ['Female', 2]], dtype=object)

By default,OrdinalEncoder is lenient towards missing values bypropagating them.

>>>importnumpyasnp>>>X=[['Male',1],['Female',3],['Female',np.nan]]>>>enc.fit_transform(X)array([[ 1.,  0.],       [ 0.,  1.],       [ 0., nan]])

You can use the parameterencoded_missing_value to encode missing values.

>>>enc.set_params(encoded_missing_value=-1).fit_transform(X)array([[ 1.,  0.],       [ 0.,  1.],       [ 0., -1.]])

Infrequent categories are enabled by settingmax_categories ormin_frequency.In the following example, “a” and “d” are considered infrequent and groupedtogether into a single category, “b” and “c” are their own categories, unknownvalues are encoded as 3 and missing values are encoded as 4.

>>>X_train=np.array(...[["a"]*5+["b"]*20+["c"]*10+["d"]*3+[np.nan]],...dtype=object).T>>>enc=OrdinalEncoder(...handle_unknown="use_encoded_value",unknown_value=3,...max_categories=3,encoded_missing_value=4)>>>_=enc.fit(X_train)>>>X_test=np.array([["a"],["b"],["c"],["d"],["e"],[np.nan]],dtype=object)>>>enc.transform(X_test)array([[2.],       [0.],       [1.],       [2.],       [3.],       [4.]])

fit(X,y=None)[source]#

Fit the OrdinalEncoder to X.

Parameters:

Xarray-like of shape (n_samples, n_features): The data to determine the categories of each feature.
yNone: Ignored. This parameter exists only for compatibility withPipeline.

Returns:

selfobject: Fitted encoder.

fit_transform(X,y=None,**fit_params)[source]#

Fit to data, then transform it.

Fits transformer toX andy with optional parametersfit_paramsand returns a transformed version ofX.

Parameters:

Xarray-like of shape (n_samples, n_features): Input samples.
yarray-like of shape (n_samples,) or (n_samples, n_outputs), default=None: Target values (None for unsupervised transformations).
**fit_paramsdict: Additional fit parameters.

Returns:

X_newndarray array of shape (n_samples, n_features_new): Transformed array.

get_feature_names_out(input_features=None)[source]#

Get output feature names for transformation.

Parameters:

input_featuresarray-like of str or None, default=None

Input features.

Ifinput_features isNone, thenfeature_names_in_ isused as feature names in. Iffeature_names_in_ is not defined,then the following input feature names are generated:["x0","x1",...,"x(n_features_in_-1)"].
Ifinput_features is an array-like, theninput_features mustmatchfeature_names_in_ iffeature_names_in_ is defined.

Returns:

feature_names_outndarray of str objects: Same as input features.

get_metadata_routing()[source]#

Get metadata routing of this object.

Please checkUser Guide on how the routingmechanism works.

Returns:

routingMetadataRequest: AMetadataRequest encapsulatingrouting information.

get_params(deep=True)[source]#

Get parameters for this estimator.

Parameters:

deepbool, default=True: If True, will return the parameters for this estimator andcontained subobjects that are estimators.

Returns:

paramsdict: Parameter names mapped to their values.

inverse_transform(X)[source]#

Convert the data back to the original representation.

Parameters:

Xarray-like of shape (n_samples, n_encoded_features): The transformed data.

Returns:

X_originalndarray of shape (n_samples, n_features): Inverse transformed array.

set_output(*,transform=None)[source]#

Set output container.

SeeIntroducing the set_output APIfor an example on how to use the API.

Parameters:

transform{“default”, “pandas”, “polars”}, default=None

Configure output oftransform andfit_transform.

"default": Default output format of a transformer
"pandas": DataFrame output
"polars": Polars output
None: Transform configuration is unchanged

Added in version 1.4:"polars" option was added.

Returns:

selfestimator instance: Estimator instance.

set_params(**params)[source]#

Set the parameters of this estimator.

The method works on simple estimators as well as on nested objects(such asPipeline). The latter haveparameters of the form<component>__<parameter> so that it’spossible to update each component of a nested object.

Parameters: