SplitStrategy

class SplitStrategy

Bases: ABC

Abstract base class for split finding strategies.

__init__()

Methods

__init__()
find_best_split(X, y[, X_val, y_val, depth])	Find the best split for the given data.
should_stop(X, y, current_gain, depth, **kwargs)	Determine if splitting should stop at this node.

abstractmethod find_best_split(X, y, X_val=None, y_val=None, depth=0, **kwargs)

Find the best split for the given data.

Parameters:

• X (ndarray) – Training feature matrix for structure learning.

• y (ndarray) – Training target values for structure learning.

• X_val (ndarray | None) – Validation feature matrix for split evaluation.

• y_val (ndarray | None) – Validation target values for split evaluation.

• depth (int) – Current depth in the tree.

• **kwargs – Strategy-specific parameters.

Returns:

Best split found, or None if no good split exists

Return type:

SplitCandidate | None

abstractmethod should_stop(X, y, current_gain, depth, **kwargs)

Determine if splitting should stop at this node.

Parameters:

• X (ndarray) – Feature matrix at current node.

• y (ndarray) – Target values at current node.

• current_gain (float) – Information gain of current best split.

• depth (int) – Current tree depth.

• **kwargs – Additional strategy-specific parameters.

Returns:

True if splitting should stop, False otherwise.

Return type:

bool