Locally linear generative topographic mapping
Abstract
We propose a method for non-linear data pro- jection that combines Generative Topographic Mapping and Coordinated PCA. We extend the Generative Topographic Mapping by using more complex nodes in the network: each node provides a linear map between the data space and the latent space. The location of a node in the data space is given by a smooth non-linear function of its location in the latent space. Our model provides a piece-wise linear mapping between data and latent space, as opposed to the point-wise coupling of the Generative Topographic Mapping. We provide experimental results comparing this model with GTM.
Domains
Machine Learning [cs.LG]
Fichier principal
verbeek02bnl.pdf (158.52 Ko)
Télécharger le fichier
VVK02g.png (10.41 Ko)
Télécharger le fichier
Origin : Files produced by the author(s)
Format : Figure, Image
Loading...