In this work, two sets of voltammetric sensors -prepared using different strategies- have been combined in an electronic tongue to evaluate the complete antioxidant profile of red wines. To this aim, wine samples were analyzed with the whole set of sensors. In order to reduce the large dimensionality of the data set while keeping the relevant information provided by the sensors, two different methods of feature selection and data compression were used (the kernels method and Discrete Wavelet Transform feature extraction method). Then, the coefficients obtained were used as the input variables of Principal Component Analysis (to evaluate the capability of discrimination. Partial-least squares regression (PLS) and artificial neural networks (ANNs) were performer to build the quantitative prediction models that allowed the quantification of the antioxidant capacity of the tested wines.
