DMCA
Developments of the generative topographic mapping (1998)
Cached
Download Links
by
Christopher M. Bishop
| Venue: | Neurocomputing |
| Citations: | 25 - 1 self |
BibTeX
@ARTICLE{Bishop98developmentsof,
author = {Christopher M. Bishop},
title = {Developments of the generative topographic mapping},
journal = {Neurocomputing},
year = {1998},
volume = {21},
pages = {203--224}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
1 Introduction Probability theory provides a powerful, consistent framework for dealing quantitatively with un-certainty (10). It is therefore ideally suited as a theoretical foundation for pattern recognition. Recently, the self-organizing map (SOM) of 19) was re-formulated within a probabilistic setting(7) to give the GTM (Generative Topographic Mapping). In going to a probabilistic formulation, several limitations of the SOM were overcome, including the absence of a cost function and thelack of a convergence proof.
