Share Email Print
cover

Proceedings Paper

Variable selection for quantitative determination of glucose concentration with near-infrared spectroscopy
Author(s): Michael J. McShane; Gerard L. Cote; Clifford H. Spiegelman
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Near-IR spectroscopy has been used in combination with multivariate calibration techniques such as partial-lest squares regression (PLSR) to quantify glucose concentration in various media. However, for reasonable prediction capability in measuring glucose many calibration samples are needed. n addition, spectroscopic data often contain over 1000 data points, presenting a very large data matrix for calibration. It is desirable to reduce the available data to contain only the information necessary for accurate prediction of chemical concentration before PLSR is applied. This will eliminate noisy variable sand consequently the data can be processed more quickly and efficiently. A variable selection method that reduces prediction bias in single factor partial least square regression models was developed and applied to near-IR absorbance spectra of glucose in two different media: pH buffer and cell culture medium. Comparisons between calibration and prediction capability for full spectra and reduced sets were completed, resulting in statistically equivalent mean squared errors. The number of response variables needed to fit the calibration data and accurately predict concentrations form new spectra was reduced in each case. The algorithm correctly chose the glucose peak areas as the informative variables and computation time was decreased by an order of magnitude.

Paper Details

Date Published: 2 May 1997
PDF: 9 pages
Proc. SPIE 2982, Optical Diagnostics of Biological Fluids and Advanced Techniques in Analytical Cytology, (2 May 1997); doi: 10.1117/12.273614
Show Author Affiliations
Michael J. McShane, Texas A&M Univ. (United States)
Gerard L. Cote, Texas A&M Univ. (United States)
Clifford H. Spiegelman, Texas A&M Univ. (United States)


Published in SPIE Proceedings Vol. 2982:
Optical Diagnostics of Biological Fluids and Advanced Techniques in Analytical Cytology
Robert C. Leif; Alexander V. Priezzhev; Toshimitsu Asakura; Robert C. Leif, Editor(s)

© SPIE. Terms of Use
Back to Top