Interference assessment and correction in the partial least squares regression method for multicomponent determination by UV spectrophotometry

Improvements have been made to the traditional partial least squares (PLS) regression and cross-validation procedures. The new PLS procedure copes more effectively with non-analyte interference in multicomponent determinations and provides an estimate of the interference spectrum. The interference spectrum can be used to partially or quantitatively correct the measured spectrum before determination of analyte concentrations. Hence the accuracy of analysis is improved. The modified cross-validation procedure reduces computer time by a factor of three, for the assessment of the number of principal components, compared with the conventional method.