WAVELENGTH SELECTION METHOD FOR MULTICOMPONENT SPECTROPHOTOMETRIC DETERMINATIONS USING PARTIAL LEAST-SQUARES

A method for eliminating unnecessary wavelengths is applied with the g oal of achieving improved prediction ability in multicomponent determi nations by UV/VIS spectrophotometry with partial least squares (PLS), The feature selection method is based on the regression coefficients o f the closed form of the PLS model, This method was evaluated with cal ibration data of different types, and with different criteria to choos e the optimum number of factors, The results presented suggest that wa velength selection improves the prediction ability of PLS method.