(39)2D correlation spectra based on higher order moments for correlation analysis with constituent concentrations
Jun Uozumi
FACSS 2010 Final Program Book of Abstracts, p.187, 2010-10
Generalized 2D correlation spectroscopy is a powerful tool for analyzing correlation properties between same or different spectra at different wavelengths. The 2D correlation spectrum of spectra A and B are defined by an average of the product of DA=A-MA and DB=B-MB with respect to an external variable t that is controlled by a certain manner, where MX stands for the mean of X. In many cases in NIR spectroscopy, 2D correlation peaks are required that reflect correlations exclusively with a certain constituent for which a calibration is to be established. Since it is not easy to control the constituent concentration as the external variable, we consider, as an alternate approach, correlations defined by higher-order moments involving DA, DB and Dp, where p is the concentration of a constituent of interest, the average being taken over many samples having different concentrations.
>>
NEXT