bnlearn: Bayesian Network Algorithms and Utility Functions

A software package by Marco Scutari of the Dalle Molle Institute for Artificial Intelligence Research implementing algorithms for Bayesian network structure learning, parameter learning, and inference. It supports constraint-based, score-based, pairwise, and hybrid learning methods for discrete, Gaussian, and conditional Gaussian networks, along with classifiers, utility functions, and estimation techniques. Development snapshots are available from the project website.

Use Cases

• Learning Bayesian network structures from data based on the described constraint-based, score-based, and hybrid algorithms.
• Estimating network parameters using maximum likelihood or Bayesian methods as mentioned in the description.
• Performing probabilistic inference and conditional probability queries using the implemented inference capabilities.
• Comparing and validating models using the included cross-validation, bootstrap, and model averaging utilities.

Strengths

• Implements a wide variety of structure learning algorithms, including PC, GS, IAMB, Hill-Climbing, Tabu Search, MMHC, and others.
• Supports multiple data distribution types: discrete, Gaussian, and conditional Gaussian networks.
• Includes comprehensive utilities for model manipulation, comparison, plotting, and validation.

Limitations

• Row count and dataset size are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

Dalle Molle Institute for Artificial Intelligence Research

Collection Method

Software package development and distribution.