NEURAL NETWORK COMPUTATIONAL PARADIGMS IN HIGH RESOLUTION SPECTROSCOPY
Loading...
Date
1989
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Ohio State University
Abstract
Neural network computational techniques1 appear to hold considerable promise for useful applications in the realm of high resolution spectroscopy. The basis for such computational models will be briefly presented followed by a selection of examples of interest. In particular, the use of backpropagation networks for resolution enhancement of spectra has been moderately successful. Examples of resolution enhancement using randomly generated network training data will be compared to ordinary (and successful) constrained deconvolution2. Resolution enhancement of FTS spectra with sinc apodization is presently under study. The status of that study also will be reported.
Description
$^{1}$ D. E. Rumelhart and J. L. McClelland, Parallel Distributed Processing, Vol. 1, MIT Press, Cambridge, Massachusetts, 1986. Neural Networks for Computing, J. S. Denker, ed., AIP Conference Proceedings 151, American Institute of Physics, New York, 1986. $^{2}$ W. E. Blass and G. W. Halsey, Deconvolution of Absorption Spectra, Academic Press, New York, 1981. Deconvolution of Spectra, P. A. Jansson, ed., Academic Press, New York, 1984.
Author Institution: Molecular Spectroscopy Laboratory, Department of Physics and Astronomy, The University of Tennessee
Author Institution: Molecular Spectroscopy Laboratory, Department of Physics and Astronomy, The University of Tennessee