Abstract
The objective of this study is to compare the effectiveness of some standard feature reduction and classification techniques (including principal component analysis, PCA; multilayer perceptrons, MLPs; and nearest neighbor classifiers, k-NN) against several proposed variants for the analysis of high-dimensional biomedical spectral data. First, the original feature space is augmented by nonlinear transformations of the original features. We present an extension of sub-pattern PCA (SpPCA) proposed by Chen, which exploits sub-patterns (rather than complete original patterns) in the process of dimensionality reduction. Comprehensive experiments demonstrate the effectiveness of SpPCA over standard PCA. SpPCA leads to the development of individually reduced subspaces and, because of the local nature of this processing, the effectiveness of the reduction and classification processes may be enhanced. With respect to classifier topologies, we present and contrast two common classifiers, MLP and k-NN, as well as several fusion strategies. Finally, some general design guidelines are offered. The experimental framework uses biomedical data acquired from magnetic resonance spectrometers.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Aleixandre M, Ares L, Fernadez MJ, Garc M, Gutirez J, Horrillo MC, Santos JP, Sayago I (2004) Analysis of neural networks and analysis of feature selection with genetic algorithm to discriminate among pollutant gas. Sens Actuators B Chem 103:122–128
Chang YL, Han CC, Jou FD, Fan KC, Chen KS, Chang JH (2004) A modular eigen subspace scheme for high-dimensional data classification. Future Gener Comp Syst 20:1131–1143
Chen S, Zhu Y (2004) Subpattern-based principal component analysis. Patt Recogn 37:1081–1083
Dalcanale E, Gardini S, Pardo M, Sberveglieri G (2000) A hierarchical classification scheme for an electronic nose. Sens Actuators B Chem 69:359–365
Duda RO, Hart PE, Stork DG (2000) Pattern classification. Wiley, New York
Gottumukkal R, Asari VK (2004) An improved face recognition technique based on modular PCA approach. Pattern Recogn Lett 25:429–436
Hughes NP, Roberts SJ, Tarassenko L (2004) Semi-supervised learning of probabilistic models for ECG segmentation. IEEE Proc EMBC 1:434–437
Kittler J, Hatef M, Duin RPW, Matas J (1998) On combining classifiers. IEEE Trans Pattern Anal Mach Intell 20:226–239
Lee YYB, Huang Y, Deredy WE, Lisboa PJG, Arus C, Harris P (2000) Robust methodology for the discrimination of brain tumours from in vivo magnetic resonance spectra. IEE Proc Sci Meas Technol 147:309–314
Lu WZ, Wang WJ, Wang XK, Yan SH, Lam JC (2004) Potential assessment of a neural network model with PCA/RBF approach for forecasting pollutant trends in Mong Kok urban air, Hong Kong. Environ Res 96:79–87
Lucht R, Delorme S, Brix G (2002) Neural network-based segmentation of dynamic MR mammographic images. Magn Reson Imag 20:147–154
Middleton I, Damper RI (2004) Segmentation of magnetic resonance images using a combination of neural networks and active contour models. Med Eng Phys 26:71–86
Mika S, Schökopf B, Smola A, Müller K-R, Scholz M, Räätsch G (1999) Kernel PCA and de-noising in feature spaces. Adv Neural Inform Process Syst 11:536–542
Mortimer RK, Contopoulou CR, King JS (1992) Genetic and physical maps of Saccharomyces cerevisiae. Yeast 8:817–902
Pang Y, Yuan Y, Li X (2007) Generalized nearest feature line for subspace learning. IEE Electron Lett 43(20):1079–1080
Pedrycz W, Breuer A, Pizzi NJ (2004) Genetic design of feature spaces for pattern classifiers. Artif Intell Med 32:115–125
Pizzi NJ (1991) Fuzzy pre-processing of gold standards as applied to biomedical spectra classification. Artif Intell Med 16:171–182
Pizzi N, Choo L-P, Mansfield J, Jackson M, Halliday WC, Mantsch HH, Somorjai RL (1995) Neural network classification of infrared spectra of control and Alzheimer’s diseased tissue. Artif Intell Med 7:67–79
Roweis S, Saul L (2000) Nonlinear dimensionality reduction by locally linear embedding. Science 290(5500):2323–2326
Ruta D, Gabrys B (2005) Classifier selection for majority voting. Inform Fusion 6:63–81
Shipp CA, Kuncheva LI (2002) Relationships between combination methods and measures of diversity in combining classifiers. Inform Fusion 3:135–148
Somorjai RL, Baumgartner R, Ho TK, Himmelreich U, Mountford CE, Sorrell TC (2003) Comparison of two strategies for classifying biomedical spectra: genetic algorithm-driven feature extraction vs. the random subspace method. Appl Environ Microbiol 69:4566–4574
Szabo BK, Aspelin P, Wiberg MK (2004) Neural network approach to the segmentation and classification of dynamic magnetic resonance images of the breast: comparison with empiric and quantitative kinetic parameters. Acad Radiol 11:1344–1354
Tang EK, Suganthan PN, Yao X, Qin AK (2005) Linear dimensionality reduction using relevance weighted LDA. Pattern Recogn 38:485–493
Turk M, Pentland A (1991) Face recognition using eigenfaces. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Lahaina, USA, June 3–5, pp 586–591
Vannucci M, Sha N, Brown PJ (2005) NIR and mass spectra classification: Bayesian methods for wavelet-based feature selection. Chemom Intell Lab Syst 77:139–148
Yang H, Irudayaraj J, Paradkar MM (2005) Discriminant analysis of edible oils and fats by FTIR, FT-NIR and FT-Raman spectroscopy. Food Chem 93:25–32
Acknowledgments
Support from the Natural Sciences and Engineering Research Council of Canada (NSERC) and the Canada Research Chair (CRC) Program (W. Pedrycz) is gratefully acknowledged.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Pedrycz, W., Lee, D.J. & Pizzi, N.J. Representation and classification of high-dimensional biomedical spectral data. Pattern Anal Applic 13, 423–436 (2010). https://doi.org/10.1007/s10044-009-0170-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10044-009-0170-1