sklearn_pmml_model.tree.tree

Module Contents

Classes

PMMLTreeClassifier	A decision tree classifier.
PMMLTreeRegressor	A decision tree regressor.

Functions

unflatten(node)	Convert a multiSplit into a binarySplit decision tree which is expressively equivalent.
construct_tree(node, classes, field_mapping[, i, ...])	Generate nodes and values used for constructing Cython Tree class.
get_tree(→ object)	Construct a single tree for a <Segment> PMML element.
clone(est[, safe])	Clone a DecisionTree, including private properties that are ignored in sklearn.base.clone.

Attributes

SPLIT_UNDEFINED

sklearn_pmml_model.tree.tree.SPLIT_UNDEFINED

class sklearn_pmml_model.tree.tree.PMMLTreeClassifier(pmml)

Bases: sklearn_pmml_model.base.PMMLBaseClassifier, sklearn.tree.DecisionTreeClassifier

A decision tree classifier.

The PMML model consists out of a <TreeModel> element, containing at least one <Node> element. Every node element contains a predicate, and optional <Node> children. Leaf nodes either have a score attribute or <ScoreDistribution> child describing the classification output.

Parameters:

pmmlstr, object

Filename or file object containing PMML data.

Notes

Specification: http://dmg.org/pmml/v4-3/TreeModel.html

fit(x, y)

Not supported: PMML models are already fitted.

_more_tags()

class sklearn_pmml_model.tree.tree.PMMLTreeRegressor(pmml)

Bases: sklearn_pmml_model.base.PMMLBaseRegressor, sklearn.tree.DecisionTreeRegressor

A decision tree regressor.

The PMML model consists out of a <TreeModel> element, containing at least one <Node> element. Every node element contains a predicate, and optional <Node> children. Leaf nodes either have a score attribute or <ScoreDistribution> child describing the classification output.

Parameters:

pmmlstr, object

Filename or file object containing PMML data.

Notes

Specification: http://dmg.org/pmml/v4-3/TreeModel.html

fit(x, y)

Not supported: PMML models are already fitted.

_more_tags()

sklearn_pmml_model.tree.tree.unflatten(node)

Convert a multiSplit into a binarySplit decision tree which is expressively equivalent.

Parameters:

nodeeTree.Element

XML Node element representing the current node.

Returns:

nodeeTree.Element

Modified XML Node element representing the flattened decision tree.

sklearn_pmml_model.tree.tree.construct_tree(node, classes, field_mapping, i=0, rescale_factor=1)

Generate nodes and values used for constructing Cython Tree class.

Parameters:

nodeeTree.Element

XML Node element representing the current node.

classeslist, None

List of possible target classes. Is None for regression trees.

field_mapping: { str: (int, callable) }

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

iint

Index of the node in the result list.

rescale_factorfloat

Factor to scale the output of every node with. Required for gradient boosting trees. Optional, and 1 by default.

Returns:

(nodes, values)tuple

nodes[()]

List of nodes represented by: left child (int), right child (int), feature (int), value (int for categorical, float for continuous), impurity (float), sample count (int) and weighted sample count (int).

values[[]]

List with training sample distributions at this node in the tree.

sklearn_pmml_model.tree.tree.get_tree(est, segment, rescale_factor=1) → object

Construct a single tree for a <Segment> PMML element.

Parameters:

est:

The estimator to built the tree for. Should contain template_estimator and field_mapping attributes.

segmenteTree.Element

<Segment> element containing the decision tree to be imported. Only segments with a <True/> predicate are supported.

rescale_factorfloat

Factor to scale the output of every node with. Required for gradient boosting trees. Optional, and 1 by default.

Returns:

treesklearn.tree.DecisionTreeClassifier, sklearn.tree.DecisionTreeRegressor

The sklearn decision tree instance imported from the provided segment, matching the type specified in est.template_estimator.

sklearn_pmml_model.tree.tree.clone(est, safe=True)

Clone a DecisionTree, including private properties that are ignored in sklearn.base.clone.

Parameters:

estBaseEstimator

The estimator or group of estimators to be cloned.

safeboolean, optional

If safe is false, clone will fall back to a deep copy on objects that are not estimators.