Events and Meetings of Italian Statistical Society, Advances in Latent Variables - Methods, Models and Applications

Font Size: 
Graphical methods for dimensionality reduction on manifolds
Lara Fontanella, Sara Fontanella, Luca Romagnoli

Last modified: 2013-06-16

Abstract


The goal of statistical methods for dimensionality reduction is to detect and discover low dimensional structures in high dimensional data. Here, we discuss a recently proposed method, known as Maximum Entropy Unfolding (MEU), for learning faithful low dimensional representations of high dimensional data. This method represents a new perspective on spectral dimensionality reduction and, joined with the theory of Gaussian Markov random fields, provides a unifying probabilistic approach to spectral dimensionality reduction techniques. Parameter estimation as well as approaches to learning the structure of the GMRF are discussed.

Full Text: PDF