sklearn_pmml_model.base

Module Contents

Classes

PMMLBaseEstimator	Base class for estimators, saving basic information on DataFields.
PMMLBaseClassifier	Base class for classifiers, preparing classes, target fields.
PMMLBaseRegressor	Base class for regressors.
OneHotEncodingMixin	Mixin class to automatically one-hot encode categorical variables.
IntegerEncodingMixin	Mixin class to automatically integer encode categorical variables.

Functions

get_type(data_field[, derives])	Parse type defined in <DataField> object and returns it.
findall(element, path)	Safe helper method to find XML elements with guaranteed return type.
parse_array(array)	Convert <Array> or <SparseArray> element into list.
parse_sparse_array(array)	Convert <SparseArray> element into list.

Attributes

array_regex

sklearn_pmml_model.base.array_regex

class sklearn_pmml_model.base.PMMLBaseEstimator(pmml)

Bases: sklearn.base.BaseEstimator

Base class for estimators, saving basic information on DataFields.

Parameters:

pmmlstr, object

Filename or file object containing PMML data.

field_mapping()

Map field name to a column index and lambda function that converts a value to the proper type.

Returns:

{ str: (int, callable) }

Dictionary mapping column names to tuples with 1) index of the column and 2) type of the column.

classmethod _get_param_names()

Get parameter names for the estimator

fields()

Return an ordered mapping from field name to XML DataField or DerivedField element.

Returns:

OrderedDict { str: eTree.Element }

Where keys indicate field names, and values are XML elements.

target_field()

Return the XML DataField or DerivedField element corresponding to the classification target.

Returns:

eTree.Element

Representing the target field for classification, or None if no MiningSchema or MiningField specified.

fit(x, y)

Not supported: PMML models are already fitted.

_prepare_data(X)

predict(X, *args, **kwargs)

Predict class or regression value for X.

This call is preceded with a data preprocessing step, that enables data scaling and categorical feature encoding.

For more information on parameters, check out the specific implementation in the scikit-learn subclass.

predict_proba(X, *args, **kwargs)

Predict class probabilities for X.

This call is preceded with a data preprocessing step, that enables data scaling and categorical feature encoding.

For more information on parameters, check out the specific implementation in the scikit-learn subclass.

sklearn_pmml_model.base.get_type(data_field, derives=None)

Parse type defined in <DataField> object and returns it.

Parameters:

data_fieldeTree.Element

<DataField> or <DerivedField> XML element that describes a column.

deriveseTree.Element

<DataField> XML element that the derived field derives.

Returns:

callable

Type of the value, as a callable function.

class sklearn_pmml_model.base.PMMLBaseClassifier(pmml)

Bases: PMMLBaseEstimator

Base class for classifiers, preparing classes, target fields.

Parameters:

pmmlstr, object

Filename or file object containing PMML data.

class sklearn_pmml_model.base.PMMLBaseRegressor(pmml)

Bases: PMMLBaseEstimator

Base class for regressors.

Parameters:

pmmlstr, object

Filename or file object containing PMML data.

sklearn_pmml_model.base.findall(element, path)

Safe helper method to find XML elements with guaranteed return type.

sklearn_pmml_model.base.parse_array(array)

Convert <Array> or <SparseArray> element into list.

Parameters:

arrayeTree.Element (Array or SparseArray)

PMML <Array> or <SparseArray> element, or type-prefixed variant (e.g., <REAL-Array>).

Returns:

outputlist

Python list containing the items described in the PMML array element.

sklearn_pmml_model.base.parse_sparse_array(array)

Convert <SparseArray> element into list.

Parameters:

arrayeTree.Element (SparseArray)

PMML <SparseArray> element, or type-prefixed variant (e.g., <REAL-SparseArray>).

Returns:

outputlist

Python list containing the items described in the PMML sparse array element.

class sklearn_pmml_model.base.OneHotEncodingMixin

Mixin class to automatically one-hot encode categorical variables.

_prepare_data(X)

class sklearn_pmml_model.base.IntegerEncodingMixin

Mixin class to automatically integer encode categorical variables.

_prepare_data(X)