19 Jun 2009
This Matlab GUI provides a simple interface for visualizing and experimenting with various dimensionality reduction and manifold learning techniques. It is intended as an instruction tool and has proven useful to researchers interested in learning about dimensionality reduction. The algorithms available include: Principal Components Analysis (PCA), Multi-Dimensional Scaling (MDS), ISOMAP, Locally Linear Embedding (LLE), Hessian eigenmaps (HLLE), Laplacian eigenmaps, Diffusion maps, and Local Tangent Space Alignment (LTSA). The GUI provides several simple datasets such as the Swiss Roll for experimentation and comparison. Users can import their own datasets from text files or from the workspace.