Classification of Spanish DO white wines according to their elemental profile by means of support vector machines.

Spanish white wines from four production areas protected by Appellation Control laws have been analysed by inductively coupled plasma optical emission spectrometry to determine the contents of aluminium, barium, boron, calcium, chromium, copper, iron, magnesium, manganese, nickel, phosphorous, potassium, silicon, sodium, strontium, sulphur and zinc. These elements were used as chemical descriptors in order to differentiate wines from different brands certified of origin. Kruskal-Wallis test was applied to highlight significant differences between the four considered classes and pattern recognition methods were applied to construct classification models. In this way, principal component analysis was used to visualise data trends and backward stepwise linear discriminant analysis was applied in order to reduce the number of input variables. The concentrations of chromium, manganese, silicon, sodium and strontium were used to construct a support vector machine classification model, obtaining a 100% of classification performance.