The Application Of Chemometrics In The Simultaneous Determination Of Multicomponent

For multicomponent systems, particularly those with similar components, an intense overlapping between the spectra and an increased complexity of samples practically involved persist in a need of useful approaches to build robust and stable linear calibration models. Partial least-squares (PLS) regression has been found useful in handling multivariate calibration in case there are many wavelengths involved and has been recommended as an approach to efficiently utilize the information carried by these wavelengths. However, there is an increasing evidence that wavelength selection is also essential for successful PLS analysis of multicomponent spectral data and the lack of variable selection also can spoil the PLS regression.It has been recognized that the elimination of uninformative spectral channels which do not contribute to model formulation is of importance even in situations where the PLS is applied.A version of modified particle swarm optimization (PSO) is proposed in the present study. This proposed method was used for simultaneous spectrophotometric determination of the content of nitrophenol isomers and cephalosporins. In comparing with PLS analysis by all wavelengths, it shows that better results are obtained from regression analysis including only selected wavelength interval than from PLS analysis including all variables. It can be seen that the PLS model using the selected wavelength interval is stable and reliable.The absorption spectra of the caffeine, phenacetin and aspirin are overlapped each other. A new method of data mining, support vector regression (SVM), has been applied to the multivariate calibration in the simultaneous determination of the complex in aqueous solution by spectrophotometric method. And the results of computation have been compared with the PLS. It has been found that the averaged accuracy of the results of SVM is better than the results obtained by PLS. So that SVM should be considered as a promising computation method in calibration of analytical chemistry.