Calibration of Artificial Neural Network for Simultaneous Spectrometric Determination of Vanillin, Ethyl vanillin, Maltol and Ethyl Maltol in Food

Artificial neural network (ANN) trained by back-propagation learning was applied to process spectrometric data for mixed maltol, ethyl maltol, vanillin and ethyl vanillin. By optimizing the network structure the accuracy was enhanced. The ANN-based method was used for the simultaneous determination of mixed maltol, ethyl maltol, vanillin and ethyl vanillin in commercial products with satisfactory results. It may also provide a new approach to spectrophotometric determination of these compounds in synthetic food without any preliminary chemical separation. The aim of this work was to apply the ANN-based methods to resolve the UV spectra of maltol, ethyl maltol, vanillin, and ethyl vanillin present as a mixture in commercial products. The proposed methods are fast, easy to apply, inexpensive, and suitable for the analysis of colorants in commercial products. Moreover, they are sensitive and selective.