CostSensitiveFS<T>
Models & Types · AiDotNet.Preprocessing.FeatureSelection.CostSensitive
Cost-Sensitive Feature Selection considering feature acquisition costs.
Inheritance: TransformerBase<T, Matrix<T>, Matrix<T>> → CostSensitiveFS<T>
Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>
For Beginners
Sometimes getting data for a feature costs money or time. This selector picks features that give you the most predictive power for the least cost. It’s like shopping for ingredients - you want quality ingredients that don’t break your budget.
How It Works
Cost-Sensitive Feature Selection balances the predictive value of features against their acquisition costs. This is important in scenarios where collecting certain features is expensive, time-consuming, or invasive.
Constructors
| Constructor | Summary |
|---|---|
CostSensitiveFS(int nFeaturesToSelect, double costWeight, Vector<T> featureCosts, int[] columnIndices) |
Properties
| Property | Summary |
|---|---|
Vector<T> CostBenefitScores { get; } |
|
Vector<T> FeatureCosts { get; } |
|
Vector<T> RelevanceScores { get; } |
|
bool SupportsInverseTransform { get; } |
|
double CostWeight { get; } |
|
int NFeaturesToSelect { get; } |
|
int[] SelectedIndices { get; } |
Methods
| Method | Summary |
|---|---|
Matrix<T> FitTransform(Matrix<T> data, Vector<T> target) |
|
T GetTotalCost() |
|
bool[] GetSupportMask() |
|
string[] GetFeatureNamesOut(string[] inputFeatureNames) |
|
void Fit(Matrix<T> data, Vector<T> target) |
|
void Fit(Matrix<T> data, Vector<T> target, Vector<T> featureCosts) |