ModifiedLLE<T>

Models & Types ยท AiDotNet.Preprocessing.DimensionalityReduction

Modified Locally Linear Embedding with regularization.

Inheritance: TransformerBase<T, Matrix<T>, Matrix<T>> โ†’ ModifiedLLE<T>

For Beginners

Modified LLE is more stable than standard LLE:

  • Works better when you have many neighbors
  • Less sensitive to noise
  • Produces more consistent results

Use cases:

  • When standard LLE is unstable
  • High number of neighbors relative to dimensions
  • Noisy data

How It Works

Modified LLE adds regularization to the standard LLE algorithm to improve numerical stability when the number of neighbors exceeds the input dimensionality.

The algorithm:

  1. Find k-nearest neighbors for each point
  2. Compute reconstruction weights with regularization
  3. Compute embedding by minimizing reconstruction error

Constructors

Constructor Summary
ModifiedLLE(int nComponents, int nNeighbors, double regParam, int? randomState, int[] columnIndices) Creates a new instance of ModifiedLLE.

Properties

Property Summary
Matrix<T> Embedding { get; } Gets the embedding result.
bool SupportsInverseTransform { get; } Gets whether this transformer supports inverse transformation.
int NComponents { get; } Gets the number of components (dimensions).
int NNeighbors { get; } Gets the number of neighbors.

Methods

Method Summary
string[] GetFeatureNamesOut(string[] inputFeatureNames) Gets the output feature names after transformation.