We introduce and apply a new classification strategy
we call computerized consensus diagnosis (CCD). Its
purpose is to provide robust, reliable classification
of biomedical data. The strategy involves the
cross-validated training of several classifiers of
diverse conceptual and methodological origin on the
same data, and appropriately combining their outcomes.
The strategy is tested on proton magnetic resonance
spectra of human thyroid biopsies, which are
successfully allocated to normal or carcinoma classes.
We used Linear Discriminant Analysis, a Neural
Net-based method, and Genetic Programming as
independent classifiers on two spectral regions, and
chose the median of the six classification outcomes as
the consensus. This procedure yielded 100% specificity
and 100% sensitivity on the training sets, and 100%
specificity and 98% sensitivity on samples of known
malignancy in the test sets. We discuss the necessary
steps any classification approach must take to
guarantee reliability, and stress the importance of
fuzziness and undecidability in robust
classification.
%0 Journal Article
%1 somorjai:1995:ccd
%A Somorjai, Ray L.
%A Nikulin, Alexander E.
%A Pizzi, Nic
%A Jackson, Dick
%A Scarth, Gordon
%A Dolenko, Brion
%A Gordon, Heather
%A Russell, Peter
%A Lean, Cynthia L.
%A Delbridge, Leigh
%A Mountford, Carolyn E.
%A Smith, Ian C. P.
%D 1995
%J Magnetic Resonance Medicine
%K ANN, GEPPETTO LDA, algorithms, classification, computerised consensus diagnosis, genetic magnetic neoplasms, programming, proton resonance robust spectrum, thyroid
%N 2
%P 257--263
%T Computerized Consensus Diagnosis: A Classification
Strategy for the Robust Analysis of MR spectra. I.
Application to 1H Spectra of Thyroid Neoplasms
%U http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=7707918&dopt=Abstract
%V 33
%X We introduce and apply a new classification strategy
we call computerized consensus diagnosis (CCD). Its
purpose is to provide robust, reliable classification
of biomedical data. The strategy involves the
cross-validated training of several classifiers of
diverse conceptual and methodological origin on the
same data, and appropriately combining their outcomes.
The strategy is tested on proton magnetic resonance
spectra of human thyroid biopsies, which are
successfully allocated to normal or carcinoma classes.
We used Linear Discriminant Analysis, a Neural
Net-based method, and Genetic Programming as
independent classifiers on two spectral regions, and
chose the median of the six classification outcomes as
the consensus. This procedure yielded 100% specificity
and 100% sensitivity on the training sets, and 100%
specificity and 98% sensitivity on samples of known
malignancy in the test sets. We discuss the necessary
steps any classification approach must take to
guarantee reliability, and stress the importance of
fuzziness and undecidability in robust
classification.
@article{somorjai:1995:ccd,
abstract = {We introduce and apply a new classification strategy
we call computerized consensus diagnosis (CCD). Its
purpose is to provide robust, reliable classification
of biomedical data. The strategy involves the
cross-validated training of several classifiers of
diverse conceptual and methodological origin on the
same data, and appropriately combining their outcomes.
The strategy is tested on proton magnetic resonance
spectra of human thyroid biopsies, which are
successfully allocated to normal or carcinoma classes.
We used Linear Discriminant Analysis, a Neural
Net-based method, and Genetic Programming as
independent classifiers on two spectral regions, and
chose the median of the six classification outcomes as
the consensus. This procedure yielded 100% specificity
and 100% sensitivity on the training sets, and 100%
specificity and 98% sensitivity on samples of known
malignancy in the test sets. We discuss the necessary
steps any classification approach must take to
guarantee reliability, and stress the importance of
fuzziness and undecidability in robust
classification.},
added-at = {2008-06-19T17:35:00.000+0200},
author = {Somorjai, Ray L. and Nikulin, Alexander E. and Pizzi, Nic and Jackson, Dick and Scarth, Gordon and Dolenko, Brion and Gordon, Heather and Russell, Peter and Lean, Cynthia L. and Delbridge, Leigh and Mountford, Carolyn E. and Smith, Ian C. P.},
biburl = {https://www.bibsonomy.org/bibtex/2e0b0fae87b2d52448eb46968284dd4a9/brazovayeye},
interhash = {f25e352d10273e1355cfc58f3755ad07},
intrahash = {e0b0fae87b2d52448eb46968284dd4a9},
issn = {0740-3194},
journal = {Magnetic Resonance Medicine},
keywords = {ANN, GEPPETTO LDA, algorithms, classification, computerised consensus diagnosis, genetic magnetic neoplasms, programming, proton resonance robust spectrum, thyroid},
month = {February},
notes = {PMID: 7707918 [PubMed - indexed for MEDLINE] consensus
means taking 2 of 3 vote from the three different
classifiers (GP, LDA and NN).},
number = 2,
pages = {257--263},
timestamp = {2008-06-19T17:51:56.000+0200},
title = {Computerized Consensus Diagnosis: {A} Classification
Strategy for the Robust Analysis of {MR} spectra. {I}.
Application to {1H} Spectra of Thyroid Neoplasms},
url = {http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=7707918&dopt=Abstract},
volume = 33,
year = 1995
}