ABSTRACT
Malinowski developed the theory of errors in abstract factor analysis and showed how experimental errors are introduced into the scheme of factor analysis. According to the concept of evolving factor analysis, the structure of spectral sets correlates with the structure of eigenvalues of the scatter matrix. According to the Lambert-Beer law for absorption spectroscopy, and according to analogous relations for other types of spectra, the increment in unit thickness is given by the product of the molar absorption coefficient and the concentration of the components. Because real spectral data are affected by experimental errors, determining the number of active components in mixture spectra is not a trivial problem. Various methods defining the primary and secondary axes of the spectral data space in a set of eigenvectors have been developed. The determination of the number of real factors (active components) starts from the lowest eigenvalue.
