rfgboost
Gradient boosting with asynchronous random forests, implemented in Rust with Python bindings.
Installation
pip install rfgboost
From source (requires Rust toolchain):
pip install maturin
maturin develop --release
Quick Start
from rfgboost import RFGBoostClassifier, RFGBoostRegressor
# Classification
clf = RFGBoostClassifier(n_estimators=20, rf_n_estimators=50, rf_max_depth=6)
clf.fit(X_train, y_train)
proba = clf.predict_proba(X_test)
ci = clf.predict_ci(X_test) # Wilson score intervals
# Regression
reg = RFGBoostRegressor(n_estimators=20, rf_n_estimators=50, rf_max_depth=6)
reg.fit(X_train, y_train)
pred = reg.predict(X_test)
ci = reg.predict_ci(X_test) # Split conformal prediction intervals
# Async mode (adaptive early stopping via CI convergence)
clf = RFGBoostClassifier(async_mode=True, tol=0.0)
# Categorical features (WOE encoding via fastwoe-rs)
clf = RFGBoostClassifier(cat_features=[0, 1, 2])
Components
| Class | Description |
|---|---|
RFGBoostClassifier |
Gradient boosting with RF base learners (binary + multiclass) |
RFGBoostRegressor |
Gradient boosting with RF base learners (regression) |
RandomForestClassifier |
Standalone random forest classifier |
RandomForestRegressor |
Standalone random forest regressor |
RandomForestUnsupervised |
Breiman's unsupervised RF (proximity, outliers, MDS) |
DecisionTree |
Single decision tree (exact sklearn match) |
TreeSHAP |
Exact tree-path-dependent SHAP values |
Key Features
- Async tree building: Rayon work-stealing with AtomicBool convergence flag. Unstarted trees skip once the ensemble converges.
- CI-based stopping: Wilson intervals (classification) and normal CI (regression) determine convergence automatically with
tol=0. - Histogram splitting: 256-bin quantile histograms for O(n + bins) split search.
- Conformal prediction: Split conformal CIs for regression with coverage guarantees.
- Unsupervised RF: Proximity matrix, outlier detection, MDS embedding, feature importance from Breiman's original method.
- Exact TreeSHAP: Matches the official SHAP package to machine precision.
License
MIT
1 maintainer
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
rfgboost-0.2.0.tar.gz
(843.2 kB
view details)
Built Distributions
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file rfgboost-0.2.0.tar.gz.
File metadata
- Download URL: rfgboost-0.2.0.tar.gz
- Upload date:
- Size: 843.2 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/6.1.0 CPython/3.13.12
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
dc26cbbcd0f27b8946af026229118dc826ced91fb930559b31d2ef4cac83d8a1
|
|
| MD5 |
86859c6c4882d27ab3d8c511db247849
|
|
| BLAKE2b-256 |
4156f5c0892852a3f6a64d3919431a2bcb9373703850118f3c8b97fd6f3d040a
|
File details
Details for the file rfgboost-0.2.0-cp39-abi3-win_amd64.whl.
File metadata
- Download URL: rfgboost-0.2.0-cp39-abi3-win_amd64.whl
- Upload date:
- Size: 455.1 kB
- Tags: CPython 3.9+, Windows x86-64
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/6.1.0 CPython/3.13.12
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
88f25b8d1720d7e11bad57037349fa87470f96cb2afa20e867be2a171f806a07
|
|
| MD5 |
1cef7c4bc3ede0901eed07c966aec545
|
|
| BLAKE2b-256 |
6ae739b65f830bd866fa9697b24cf0e2177c82b44ef717074abec6786ab40526
|
File details
Details for the file rfgboost-0.2.0-cp39-abi3-pyemscripten_2025_0_wasm32.whl.
File metadata
- Download URL: rfgboost-0.2.0-cp39-abi3-pyemscripten_2025_0_wasm32.whl
- Upload date:
- Size: 215.3 kB
- Tags: CPython 3.9+, PyEmscripten 2025.0 wasm32
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/6.1.0 CPython/3.13.12
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
e3a06d63628cec3fb5d8d576259b5801015a0fabb2b16984b6c2c25ab76a7e63
|
|
| MD5 |
47134a447f0265b048a091938f22d89c
|
|
| BLAKE2b-256 |
3866e09de23e4e2adc81a81c3e50e89f17d16d49ac75aeba88c7f6a0efc21c70
|
File details
Details for the file rfgboost-0.2.0-cp39-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.
File metadata
- Download URL: rfgboost-0.2.0-cp39-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
- Upload date:
- Size: 634.7 kB
- Tags: CPython 3.9+, manylinux: glibc 2.17+ x86-64
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/6.1.0 CPython/3.13.12
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
366c3f9d027591a0db03f608e054f677588c4fdc1e29a334738b41a9009e9cae
|
|
| MD5 |
441e4f8a9fb4a6e101ee322452a68153
|
|
| BLAKE2b-256 |
d4f44dd507b77511bb5d49ddea1e2110cc448dbd0eec4cc7e390194450c775aa
|
File details
Details for the file rfgboost-0.2.0-cp39-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl.
File metadata
- Download URL: rfgboost-0.2.0-cp39-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
- Upload date:
- Size: 619.8 kB
- Tags: CPython 3.9+, manylinux: glibc 2.17+ ARM64
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/6.1.0 CPython/3.13.12
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
ff79b1ba23d7b34fcea678d48b6e117f8033c6058024edc0f8b97c282f67ad3d
|
|
| MD5 |
21daaebc0fb2183907fe957aa1594d8b
|
|
| BLAKE2b-256 |
71bd54e16a75cd33c6cedbf7353008321c30381403a5f88f30c3823c8b4dda56
|
File details
Details for the file rfgboost-0.2.0-cp39-abi3-macosx_11_0_arm64.whl.
File metadata
- Download URL: rfgboost-0.2.0-cp39-abi3-macosx_11_0_arm64.whl
- Upload date:
- Size: 557.4 kB
- Tags: CPython 3.9+, macOS 11.0+ ARM64
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/6.1.0 CPython/3.13.12
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
75970f0c380d4f4a821d752c4b233823d2c301630a39c480f82af649fe693612
|
|
| MD5 |
2bfe605d2cccf92a792eadb5f059d266
|
|
| BLAKE2b-256 |
c7468956dbc37b33c7567f204fc3bfcbec078bfab68383765df5b4f76a082c94
|
File details
Details for the file rfgboost-0.2.0-cp39-abi3-macosx_10_12_x86_64.whl.
File metadata
- Download URL: rfgboost-0.2.0-cp39-abi3-macosx_10_12_x86_64.whl
- Upload date:
- Size: 568.4 kB
- Tags: CPython 3.9+, macOS 10.12+ x86-64
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/6.1.0 CPython/3.13.12
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
7527d7a2b32941572349a9d00a4cf217db054165251f48f16785ed171fcf6006
|
|
| MD5 |
7cc5a7407b0193d9d0d340ded9bcf932
|
|
| BLAKE2b-256 |
a2e63b968d063102c25e113bd18303a52799774286f872de331dde0065f2a969
|
