New preprocessing methods for better classification of MR and IR spectra

We introduce a global feature extraction method specifically designed to preprocess magnetic resonance spectra of biomedical origin. Such preprocessing is essential for the accurate and reliable classification of diseases or disease stages manifest in the spectra. The new method is Genetic Algorithm...

Full description

Bibliographic Details
Main Author:	Nikouline, Alexandre
Language:	en_US
Published:	2007
Online Access:	http://hdl.handle.net/1993/1521

id	ndltd-MANITOBA-oai-mspace.lib.umanitoba.ca-1993-1521
record_format	oai_dc
spelling	ndltd-MANITOBA-oai-mspace.lib.umanitoba.ca-1993-15212014-01-31T03:30:45Z New preprocessing methods for better classification of MR and IR spectra Nikouline, Alexandre We introduce a global feature extraction method specifically designed to preprocess magnetic resonance spectra of biomedical origin. Such preprocessing is essential for the accurate and reliable classification of diseases or disease stages manifest in the spectra. The new method is Genetic Algorithm-guided. It is compared with our enhanced version of the Forward Selection algorithm ("Dynamic Programming"). Both seek and select optimal spectral subregions. These subregions necessarily retain spectral information, thus aiding the eventual identification of the biochemistry of disease presence and progression. Both methods proved to be very useful for large datasets. The danger of overfitting related to the small number of samples in the datasets was demonstrated for both the artificial and real-life data. A bilinear regression model was used to quantitate the consequences of overfitting. Taking this in account, optimal parameters for the GA guided algorithm were recommended. 2007-05-17T12:39:15Z 2007-05-17T12:39:15Z 1998-03-01T00:00:00Z http://hdl.handle.net/1993/1521 en_US
collection	NDLTD
language	en_US
sources	NDLTD
description	We introduce a global feature extraction method specifically designed to preprocess magnetic resonance spectra of biomedical origin. Such preprocessing is essential for the accurate and reliable classification of diseases or disease stages manifest in the spectra. The new method is Genetic Algorithm-guided. It is compared with our enhanced version of the Forward Selection algorithm ("Dynamic Programming"). Both seek and select optimal spectral subregions. These subregions necessarily retain spectral information, thus aiding the eventual identification of the biochemistry of disease presence and progression. Both methods proved to be very useful for large datasets. The danger of overfitting related to the small number of samples in the datasets was demonstrated for both the artificial and real-life data. A bilinear regression model was used to quantitate the consequences of overfitting. Taking this in account, optimal parameters for the GA guided algorithm were recommended.
author	Nikouline, Alexandre
spellingShingle	Nikouline, Alexandre New preprocessing methods for better classification of MR and IR spectra
author_facet	Nikouline, Alexandre
author_sort	Nikouline, Alexandre
title	New preprocessing methods for better classification of MR and IR spectra
title_short	New preprocessing methods for better classification of MR and IR spectra
title_full	New preprocessing methods for better classification of MR and IR spectra
title_fullStr	New preprocessing methods for better classification of MR and IR spectra
title_full_unstemmed	New preprocessing methods for better classification of MR and IR spectra
title_sort	new preprocessing methods for better classification of mr and ir spectra
publishDate	2007
url	http://hdl.handle.net/1993/1521
work_keys_str_mv	AT nikoulinealexandre newpreprocessingmethodsforbetterclassificationofmrandirspectra
_version_	1716627998335041536

New preprocessing methods for better classification of MR and IR spectra

Similar Items