Manifold Learning

Dimensionality reduction technique that seeks to represent complex data in a lower-dimensional space, while maintaining their geometric structure.