PCAlgorithmSelector<T>
Models & Types ยท AiDotNet.Preprocessing.FeatureSelection.Causal
PC Algorithm-based Feature Selection for causal discovery.
Inheritance: TransformerBase<T, Matrix<T>, Matrix<T>> โ PCAlgorithmSelector<T>
Implements: ISupervisedFeatureSelector<T>
For Beginners
The PC algorithm tries to figure out what causes what by testing if variables become independent when controlling for other variables. Features that are directly connected to the target in this causal web are the most important for understanding and prediction.
How It Works
Uses the PC (Peter-Clark) algorithm to learn a causal graph structure from data, then selects features that are directly connected to the target variable in the causal graph.
Constructors
| Constructor | Summary |
|---|---|
PCAlgorithmSelector(int maxFeatures, double alpha, int maxConditioningSetSize, int[] columnIndices) |
Properties
| Property | Summary |
|---|---|
bool SupportsInverseTransform { get; } |
|
bool[] AdjacencyMatrix { get; } |
|
double Alpha { get; } |
|
double[] ConnectionStrengths { get; } |
|
int MaxFeatures { get; } |
|
int[] SelectedIndices { get; } |
Methods
| Method | Summary |
|---|---|
Matrix<T> FitTransform(Matrix<T> data, Vector<T> target) |
|
bool[] GetSupportMask() |
|
string[] GetFeatureNamesOut(string[] inputFeatureNames) |
|
void Fit(Matrix<T> data, Vector<T> target) |