scikit-bayes

scikit-bayes is a Python package that extends scikit-learn with a suite of Bayesian Network Classifiers.

The primary goal of this package is to provide robust, scikit-learn-compatible implementations of advanced Bayesian classifiers that are not available in the core library.

Key Features

• MixedNB: Naive Bayes for mixed data types (Gaussian + Categorical + Bernoulli) in a single model
• AnDE: Averaged n-Dependence Estimators (AODE, A2DE) that relax the independence assumption
• ALR: Accelerated Logistic Regression - hybrid generative-discriminative models with 4 weight granularity levels
• WeightedAnDE: Discriminatively-weighted ensemble models
• Full scikit-learn API: Compatible with pipelines, cross-validation, and grid search

Quick Start

import numpy as np
from skbn import MixedNB, AnDE

# MixedNB: Handle mixed data types automatically
X = np.array([[1.5, 0, 2], [-0.5, 1, 0], [2.1, 1, 1], [-1.2, 0, 2]])
y = np.array([0, 1, 1, 0])

clf = MixedNB()
clf.fit(X, y)
print(clf.predict([[0.5, 1, 1]]))  # Automatically handles Gaussian, Bernoulli, Categorical

# AnDE: Solve problems Naive Bayes cannot (XOR)
X_xor = np.array([[-1, -1], [-1, 1], [1, -1], [1, 1]])
y_xor = np.array([0, 1, 1, 0])

clf = AnDE(n_dependence=1, n_bins=2)
clf.fit(X_xor, y_xor)
print(clf.predict(X_xor))  # [0, 1, 1, 0] ✓

Installation

pip install scikit-bayes

Or install from source:

pip install git+https://github.com/ptorrijos99/scikit-bayes.git

Documentation

• 📖 User Guide - Detailed documentation
• 📚 API Reference - Complete API docs
• 🎨 Examples Gallery - Visual examples

Development

This project uses pixi for environment management.

# Run tests
pixi run test

# Run linter
pixi run lint

# Build documentation
pixi run build-doc

# Activate development environment
pixi shell -e dev

Citation

If you use scikit-bayes in a scientific publication, please cite:

@software{scikit_bayes,
  author = {Torrijos, Pablo},
  title = {scikit-bayes: Bayesian Network Classifiers for Python},
  year = {2025},
  url = {https://github.com/ptorrijos99/scikit-bayes}
}

References

• Webb, G. I., Boughton, J., & Wang, Z. (2005). Not so naive Bayes: Aggregating one-dependence estimators. Machine Learning, 58(1), 5-24.
• Flores, M. J., Gámez, J. A., Martínez, A. M., & Puerta, J. M. (2009). GAODE and HAODE: Two proposals based on AODE to deal with continuous variables. ICML '09, 313-320.
• Zaidi, N. A., Webb, G. I., Carman, M. J., & Petitjean, F. (2017). Efficient parameter learning of Bayesian network classifiers. Machine Learning, 106(9-10), 1289-1329.

License

BSD-3-Clause. See LICENSE for details.