Instrumental effects in the principal component structure of a set of spectra

Principal components analysis (PCA) is a mathematical technique widely used for the diagnosis of structure in a set of spectra, followed by factor analysis for decomposition of the spectra into physically meaningful constituents. Among the effects that can influence the principal component structure for a set of spectra, one can distinguish effects that come from the imperfect operation of particular devices that constitute the spectrometer. When one wants to apply PCA to detect subtle effects (e.g. very small shifts between spectral lines), then instrumental imperfections disturbing the data cannot be neglected, even if they are not very distinct. This paper reviews, classifies and presents the influence of instrumental effects on the principal component structure of recorded spectra. Copyright © 2002 John Wiley & Sons, Ltd.