BayesianCausalBase<T>
Base Classes · AiDotNet.CausalDiscovery.Bayesian
Base class for Bayesian causal discovery algorithms.
Inheritance: CausalDiscoveryBase<T> → BayesianCausalBase<T>
Known derived types: BCDNetsAlgorithm<T>, BayesDAGAlgorithm<T>, DiBSAlgorithm<T>, IterativeMCMCAlgorithm<T>, OrderMCMCAlgorithm<T>, PartitionMCMCAlgorithm<T>
For Beginners
Instead of returning a single “best guess” graph, Bayesian methods give probabilities for each possible connection. This tells you not just “X probably causes Y” but also “we’re 90% confident about this.” The trade-off is higher computation.
How It Works
Bayesian methods maintain a posterior distribution over possible DAG structures given the data. They can represent uncertainty about the causal structure and are naturally suited for model averaging. Methods include MCMC sampling over graphs, variational inference, and gradient-based approaches.
Properties
| Property | Summary |
|---|---|
CausalDiscoveryCategory Category { get; } |