HillClimbingAlgorithm<T>
Models & Types · AiDotNet.CausalDiscovery.ScoreBased
Hill Climbing — greedy score-based DAG structure learning.
Inheritance: CausalDiscoveryBase<T> → ScoreBasedBase<T> → HillClimbingAlgorithm<T>
For Beginners
Hill climbing is the simplest score-based approach. At each step, it tries adding, removing, or flipping every possible edge and picks the change that improves the score the most.
How It Works
Hill climbing searches over individual DAGs by greedily applying the single-edge operation (add, remove, or reverse) that most improves the BIC score.
Reference: Heckerman et al. (1995), “Learning Bayesian Networks”.
Constructors
| Constructor | Summary |
|---|---|
HillClimbingAlgorithm(CausalDiscoveryOptions options) |
Properties
| Property | Summary |
|---|---|
bool SupportsNonlinear { get; } |
|
string Name { get; } |