Trees

The Trees module extracts branches and rules from a trained forest. It supports scikit-learn and PySpark models.

Signature

Trees(
    backend="sklearn",
    n_sample_multiplier=0.05,
    ef_sample_multiplier=10,
    max_depth=None,
)

Main functions

• get_branches(X, y, model=None): returns the forest branches. When model is not provided, the forest is trained internally.
• plot_importances(model): generates an importance plot for quick diagnostics.

Typical usage

from InsideForest import Trees

trees = Trees("sklearn", n_sample_multiplier=0.05, ef_sample_multiplier=10)
branches = trees.get_branches(X, y, fitted_model)

Key parameters

• backend: "sklearn" or "pyspark" depending on the forest type.
• n_sample_multiplier: controls the sample used to evaluate rules.
• ef_sample_multiplier: adjusts exploratory sampling to discover additional regions.
• max_depth: limits depth while extracting rules.

The output feeds into Regions and Labels. See Performance Tips for parameter guidance.