Using Artificial Neural Network for Simultaneous Spectrophotometric Determination of Cobalt and Nickel

Abstract

Simultaneous spectrophotometric determination of cobalt(II) and nickel(II) based on formation of their complexes with ethylenediamine tetraacetic acid (EDTA) has been studied. An artificial neural network (ANN) model has been used to analyze the mixture spectra. A series of standard solutions of Co-EDTA and Ni-EDTA as calibration set and testing set were prepared and the absorbance spectra in the wavelength between 530-690 nm have been recorded. Calibration set for training and optimization parameters of ANN such as learning rate, number of neurons in the hidden layer and iteration have been used. The performance of ANN has been evaluated with testing set and then in all steps the root mean square errors (RMSE) were measured. The RMSE for cobalt and nickel are 0.017 and 0.024, respectively.