bartz.grove

Functions to create, manipulate, and check binary decision trees.

Tree representation

TreeHeaps(*args, **kwargs)	A protocol for dataclasses that represent trees.
TreesTrace(var_tree, split_tree, leaf_tree)	Implementation of bartz.grove.TreeHeaps for an MCMC trace.
is_multivariate(trees)	Return whether the trees have vector-valued leaves.

Evaluation

evaluate_forest(X, trees, *[, sum_batch_axis])	Evaluate an ensemble of trees at an array of points.
traverse_tree(x, var_tree, split_tree)	Find the leaf where a point falls into.
traverse_forest(X, var_trees, split_trees)	Find the leaves where points falls into for each tree in a set.

Tree and node properties

is_actual_leaf(split_tree, *[, add_bottom_level])	Return a mask indicating the leaf nodes in a tree.
is_leaves_parent(split_tree)	Return a mask indicating the nodes with leaf (and only leaf) children.
tree_depth(tree)	Return the maximum depth of a tree.
tree_actual_depth(split_tree)	Measure the depth of the tree.
tree_depths(tree_size)	Return the depth of each node in a binary tree.

Forest summaries

forest_mean_leaves(split_tree)	Return the average number of leaves per tree in a set of trees.
forest_depth_distr(split_tree)	Histogram the depths of a set of trees.
points_per_node_distr(X, var_tree, ...[, ...])	Histogram points-per-node counts in a set of trees.
var_histogram(p, var_tree, split_tree, *[, ...])	Count how many times each variable appears in a tree.

Validation and inspection

check_trace(trace, max_split)	Check the validity of a set of trees.
describe_error(error)	Describe an error code returned by check_trace.
format_tree(tree, *[, print_all])	Convert a tree to a human-readable string.