MaximalInformation<T>
Models & Types · AiDotNet.Preprocessing.FeatureSelection.Unsupervised
Maximal Information feature selection based on entropy for unsupervised learning.
Inheritance: TransformerBase<T, Matrix<T>, Matrix<T>> → MaximalInformation<T>
Implements: IFixedCountFeatureSelector<T>
For Beginners
Entropy measures how “surprising” or “uncertain” a feature’s values are. A feature that’s always the same is boring (low entropy). A feature with lots of different values is informative (high entropy). This method selects the most informative features.
How It Works
Maximal Information selects features with highest entropy (information content). Features with higher entropy contain more information and can better capture the underlying structure of the data.
Constructors
| Constructor | Summary |
|---|---|
MaximalInformation(int nFeaturesToSelect, int nBins, int[] columnIndices) |
Properties
| Property | Summary |
|---|---|
Vector<T> Entropies { get; } |
|
bool SupportsInverseTransform { get; } |
|
int NBins { get; } |
|
int NFeaturesToSelect { get; } |
|
int[] SelectedIndices { get; } |
Methods
| Method | Summary |
|---|---|
bool[] GetSupportMask() |
|
string[] GetFeatureNamesOut(string[] inputFeatureNames) |