Principal component analysis (PCA) has been used for quantitative analysis of infrared attenuated total reflection (ATR) spectra of chemical mixtures, so that the minute concentration of glucose in the chemical mixture would be evaluated. Spectroscopic calibration of chemical concentration has been, thus far, believed to require multidimensional least-squares methods, such as classical least squares (CLS) and partial least-squares (PLS) regression techniques. In the present study, however, the analysis of the infrared ATR spectra only by PCA has been found to be useful for the evaluation of concentration of specific minute species in a chemical mixture. The prediction accuracy by PCA only has proved to be comparable to that of PLS. The novel technique would enable us to measure the concentration of glucose in human blood with no invasion.
|Japanese Journal of Applied Physics, Part 2: Letters
|Published - 2002 4月 15
ASJC Scopus subject areas