Abstract: The partial least-square (PLS) and multivariable linear regression methods are used in the simultaneous determination of Co, Ni and Cu mixtures. The effects of parameters, the number of calibration samples n, the number of wavelengths p and the latent variables d, on the prediction of results are discussed. The results show that the accuracy of PLSanalysis is superior to that of MR method for determining a multicomponent system in the case of the spectra with a low selectivity.

Key words: Partial least-square method, Multivariable linear regression method, Simultaneous determination of Co, Ni and Cu