غلامرضا عبدی

A multicomponent analysis method for the simultaneous spectrophotometric determination of the Cd2+, Cu2+, and Zn2+ based on complex formation with dimethyl-spiro[isobenzofurane-1,6′-pyrorolo[2,3-d]pyrimidine]-2′,3,4,5′(1′H,3′H,7′H)tetraone using wavelet transformation–feed forward neural network is proposed. The analytical data showed that metal to ligand ratios in all metal complexes was 1:1. The absorption spectra were evaluated with respect to synthetic ligand concentration and pH. It was found that, at pH 6.7, the complexation reactions were completed. Spectral data were reduced using continuous wavelet transformation (CWT) and subjected to artificial neural networks. The presence of nonlinearities was confirmed by a partial response plot. The structures of the CWT–feed forward neural networks (WT-FFNN) were simplified using the corresponding wavelet coefficients of mother wavelets. Once the optimal wavelet coefficients are selected, different ANN models can be employed for the calculation of the final calibration model. The proposed methods were successfully applied to the simultaneous determination of Cd2+, Cu2+, and Zn2+ in rice, dill, tomato, and lettuce samples.