Application of Wavelet Transform-Principal Component Regression Method for Simultaneous Determination of Colourants in Ternary Mixtures Using Spectrophotometry

Abstract

A new multivariate method combined wavelet transform and principal component regression techniques for the simultaneous determination of safranine, phloxine B and eosinY using total absorbance measurements is proposed. In this method, the visible absorption spectra of ternary mixtures was pretreated by wavelet transform at first,and then the principal component regression procedure was performed on the wavelet coefficients of the original spectra. The absorption spectra and related wavelet coefficients (both approximation and detail) at different scales were used to perform the optimization of the calibration matrices by the principal component regression method for a comparative study, it showed that the model based on wavelet approach coefficients at lower scale is better than that based on the full visible absorption spectra. A principal component regression multivariate calibration model on the first approach coefficients of the original visible absorption spectra extracted by Bior1.1 wavelet transform, using cross-validation method leaving one out calibration sample at a time to select the optimum number of components, produced a satisfactory result with good prediction accuracies.