scikit-bayes#

Date: Feb 19, 2026 Version: 0.1.4

Useful links: Source Repository | Issues & Ideas |

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

  • AnDE Family: Averaged n-Dependence Estimators that relax the independence assumption

  • Hybrid Models: ALR and WeightedAnDE for discriminative learning

Getting started

Information about installation and basic usage of scikit-bayes.

User guide

Learn how to create your own scikit-learn compatible estimators.

API reference

Complete reference documentation for all estimators and utilities.

Examples

Gallery of examples demonstrating AnDE, MixedNB, and more.