2 Minimax theorem for graph embeddings

  • MitsubishiElectricResearchLabs CambridgeMA
  • Published 2004
Spectral methods for nonlinear dimensionality reduction (NLDR) impose a neighborhood graph on point data and compute eigenfunctions of a quadratic form generated from the graph. We introduce a more general and more robust formulation of NLDR based on the singular value decomposition (SVD). In this framework, most spectral NLDR principles can be recovered by… CONTINUE READING