Karhunen-Loeve (PCA) components of a Gaussian Process: Matlab example (2), animation and posterior

Antonio Sala, UPV

Difficulty: **** ,       Relevance: ,      Duration: 16:44

Summary:

This video is a continuation of the video [gpkh2EN], where the meaning of a matrix square root of the covariance $K=Q{Q}^T$ and its continuous limit (Karhunen-Loève eigenfunctions) were discussed. In this video, realizations are generated by adding component by component multiplied by a standard normal variable, in a Matlab animation. The second part of the video computes the posterior assuming that the process value has been observed at some points (see video [gpsambpoEN] for details). Then, diagonalizing the covariance, the shape of the principal components (scaled eigenvectors) is analyzed and realizations of the posterior are constructed from them.

*Link to my [ whole collection] of videos in English. Link to larger [ Colección completa] in Spanish.