neural_trees.MultivariateDecisionTree#

class neural_trees.MultivariateDecisionTree(max_depth: int = 4, min_samples_split: int = 10, min_samples_leaf: int = 3, min_impurity_decrease: float = 0.0, random_state: int | None = None, class_weight=None, min_weight_fraction_leaf: float = 0.0)[source]#

Bases: ClassifierMixin, BaseEstimator

Multivariate Decision Tree Classifier (sklearn-compatible).

Each internal node splits on a linear combination of all features, w . x + b > 0, where w is a linear discriminant fitted on the samples reaching that node. Boundaries are oblique rather than axis-aligned, so correlated features are handled in one node instead of a staircase of univariate cuts.

Parameters:
max_depthint, default=4

Maximum depth of the tree.

min_samples_splitint, default=10

Minimum samples required to attempt a split at a node.

min_samples_leafint, default=3

Minimum samples that must land on each side of a split.

min_impurity_decreasefloat, default=0.0

Minimum weighted Gini decrease required to keep a split.

min_weight_fraction_leaffloat, default=0.0

Smallest fraction of the total sample weight a child may hold, as in scikit-learn’s trees. Only meaningful with sample_weight or class_weight; it stops a split from isolating a region whose weight is negligible, which is how a down-weighted class would otherwise keep leaves of its own.

random_stateint or None, default=None

Seed for the centroid clustering used to build two-group problems when a node holds more than two classes.

class_weightdict, “balanced” or None, default=None

Weights per class, combined multiplicatively with sample_weight. “balanced” uses n_samples / (n_classes * bincount(y)).

Attributes:
classes_ndarray of shape (n_classes,)
n_features_in_int
root_internal node object
tree_depth_int

Depth actually reached after fitting.

n_nodes_int

Number of internal (splitting) nodes.

References

Alpaydın, E., & Çetin, Ü. (1995). Multivariate Statistical Techniques for Constructive Induction. Yıldız, O. T., & Alpaydın, E. (2001). Omnivariate Decision Trees. IEEE TNN.

Examples

>>> from neural_trees import MultivariateDecisionTree
>>> from sklearn.datasets import load_wine
>>> X, y = load_wine(return_X_y=True)
>>> mdt = MultivariateDecisionTree(max_depth=3, random_state=0)
>>> mdt.fit(X, y)
>>> mdt.score(X, y)
fit(X, y, sample_weight=None) → MultivariateDecisionTree[source]#

Fit the multivariate tree.

Parameters:
Xarray-like of shape (n_samples, n_features)
yarray-like of shape (n_samples,)
sample_weightarray-like of shape (n_samples,), default=None

Enters the discriminant at every node (weighted class means, pooled covariance and prior ratio), the Gini decrease and the leaf distributions. min_samples_split and min_samples_leaf keep counting rows, as scikit-learn’s trees do, so weighting a row is not the same as repeating it where those limits bind. Without weights the node discriminant is scikit-learn’s own solver, unchanged.

Returns:
self
get_split_weights() → List[Tuple[ndarray, float]][source]#

Return the hyperplane of every internal node as (w, b) pairs, in pre-order. Useful for reading off which features drive a split.

predict(X) → ndarray[source]#

Predict class labels.

Parameters:
Xarray-like of shape (n_samples, n_features)
Returns:
y_predndarray of shape (n_samples,)
predict_proba(X) → ndarray[source]#

Predict class probabilities from the reached leaf’s class distribution.

Parameters:
Xarray-like of shape (n_samples, n_features)
Returns:
probandarray of shape (n_samples, n_classes)
set_fit_request(*, sample_weight: bool | None | str = '$UNCHANGED$') → MultivariateDecisionTree#

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

  • True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

  • False: metadata is not requested and the meta-estimator will not pass it to fit.

  • None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

  • str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:
sample_weightstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED

Metadata routing for sample_weight parameter in fit.

Returns:
selfobject

The updated object.

set_score_request(*, sample_weight: bool | None | str = '$UNCHANGED$') → MultivariateDecisionTree#

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

  • True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

  • False: metadata is not requested and the meta-estimator will not pass it to score.

  • None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

  • str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:
sample_weightstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED

Metadata routing for sample_weight parameter in score.

Returns:
selfobject

The updated object.