DocumentCode :
419554
Title :
Consensus-based identification of spectral signatures for classification of high-dimensional biomedical spectra
Author :
Pranckeviciene, Erinija ; Baumgartner, Richard ; Somorjai, Ray
Author_Institution :
Inst. for Biodiagnostics, NRCC, Winnipeg, Man., Canada
Volume :
2
fYear :
2004
fDate :
23-26 Aug. 2004
Firstpage :
319
Abstract :
The identification of spectral signatures is crucial for the classification/profiling of biomedical spectra. Because only limited number of biomedical samples of high dimensionality is typically available, dimensionality reduction techniques (identification of discriminatory features) are essential for robust classifier development. We show, on three real-world biomedical datasets, the potential of a consensus-based identification of important feature subsets, using a genetic algorithm and a sparse linear classifier. When training data are in short supply, the proposed methodology leads to more stable subset identification and higher classification accuracy.
Keywords :
feature extraction; genetic algorithms; identification; image classification; medical image processing; biomedical datasets; biomedical spectra classification; consensus identification; discriminatory feature identification; genetic algorithm; sparse linear classifier; spectral signatures; Biomarkers; Data mining; Electronic mail; Feature extraction; Genetic algorithms; Linear programming; Mathematical programming; Principal component analysis; Robustness; Training data;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Pattern Recognition, 2004. ICPR 2004. Proceedings of the 17th International Conference on
ISSN :
1051-4651
Print_ISBN :
0-7695-2128-2
Type :
conf
DOI :
10.1109/ICPR.2004.1334189
Filename :
1334189
Link To Document :
بازگشت