TSNE<T>

Models & Types ยท AiDotNet.Preprocessing.DimensionalityReduction

t-Distributed Stochastic Neighbor Embedding for visualization.

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

For Beginners

t-SNE creates beautiful 2D/3D visualizations:

  • Points that are similar stay close together
  • Points that are different move apart
  • Great for exploring clusters in your data
  • Warning: Not for preserving global distances, just local neighborhoods
  • Warning: Results can vary with different random seeds

How It Works

t-SNE is a nonlinear dimensionality reduction technique well-suited for visualizing high-dimensional data in 2D or 3D space. It preserves local structure by keeping similar points close together.

The algorithm converts similarities between data points to joint probabilities and tries to minimize the divergence between probability distributions in high and low dimensional spaces.

Constructors

Constructor Summary
TSNE(int nComponents, double perplexity, double learningRate, int nIter, double earlyExaggeration, TSNEMetric metric, TSNEInitialization initialization, int? randomState, int[] columnIndices) Creates a new instance of TSNE.

Properties

Property Summary
Matrix<T> Embedding { get; } Gets the embedding result.
TSNEMetric Metric { get; } Gets the distance metric.
bool SupportsInverseTransform { get; } Gets whether this transformer supports inverse transformation.
double LearningRate { get; } Gets the learning rate.
double Perplexity { get; } Gets the perplexity parameter.
int NComponents { get; } Gets the number of components (dimensions).

Methods

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