Registry / ai-ml / quantile-forest

quantile-forest

library1.4.1pypypi✓ verified 82d ago

Quantile regression forests for scikit-learn compatible probabilistic forecasting. Version 1.4.1, actively maintained. Provides quantile predictions from random forest models with scikit-learn API.

ai-ml

pip install quantile-forest

github ↗homepage ↗

INSTALL

IMPORT

SIG · QUANTILE-FOREST

quantile-forest

ai-mlpythonv1.4.1

harness data pending

Install & Compatibility

Where this runs

No compatibility data collected yet for this library.

Code

Verified usage

Verified import paths — ran on the pinned version, not inferred.

RandomForestQuantileRegressor

✓ from quantile_forest import RandomForestQuantileRegressor

✗ from quantile_forest import QuantileForestRegressor

ExtraTreesQuantileRegressor

✓ from quantile_forest import ExtraTreesQuantileRegressor

Fit a quantile regression forest and predict 5th and 95th percentiles.

from quantile_forest import QuantileForestRegressor from sklearn.datasets import make_regression X, y = make_regression(n_samples=100, n_features=4, random_state=0) qfr = QuantileForestRegressor() qfr.fit(X, y) y_pred_lower = qfr.predict(X, quantiles=0.05) y_pred_upper = qfr.predict(X, quantiles=0.95) print(y_pred_lower[:5], y_pred_upper[:5])

Debug

Known issues

gotchapredict() returns numpy array with shape (n_samples, n_quantiles) when multiple quantiles are provided, but (n_samples,) for single quantile. Check shape assumptions.

fix

Always use quantiles as a list even for single value: quantiles=[0.5] to get consistent shape.

affects: all

gotchaThe fit() method requires y to be 1-dimensional. Passing a 2D array (e.g., shape (n,1)) may fail silently or raise error.

fix

Squeeze y: y = y.ravel() before fitting.

affects: all

deprecatedIn version 1.3.0, the parameter 'n_estimators' was renamed to 'n_estimators' (no change), but 'bootstrap' param default changed from False to True. Re-check your model config if upgrading.

fix

Explicitly set bootstrap=False if you relied on old default.

affects: >=1.3.0

gotchaWhen using predict() with return_y_std=True, the output is a tuple of (y_hat, y_std) but only for the mean prediction (quantiles=None). Do not combine with quantiles.

fix

Use quantiles=None to get mean and std; use quantiles=[0.5] to get median separately.

affects: all

Upgrade

Version history

1.4.1latest on PyPI · released Sep 10, 2025

Audit

Dependencies

scikit-learnrequiredCore dependency, provides base estimators.

numpyrequiredNumerical operations.

pandasoptionalOptional, for DataFrame inputs.

Agent activity

10 hits · last 30 days

node

OpenAI (training)

Resources

githubhttps://github.com/zillow/quantile-forest ↗

homepagehttps://github.com/zillow/quantile-forest ↗