Recognition of the geographical origin of beer based on support vector machines applied to chemical descriptors

Abstract A study on the differentiation of the geographical provenance of European beers from Germany, Portugal and Spain, has been carried out. Several chemical descriptors have been selected according to their significance on the brewing process as well as their use as quality indicators. Contents of aluminium, barium, boron, calcium, iron, magnesium, manganese, phosphorus, potassium, sodium, strontium, zinc, chloride, phosphate, sulphate, total amino acids, total polyphenols, pH, real extract and absorbance at 430 nm have been determined in beer samples. Kruskal–Wallis test was used to find out statistical differences among the considered origins. Backward stepwise linear discriminant analysis was applied in order to reduce the number of chemical descriptors to be considered. The selected variables were the contents in iron, phosphorus, potassium, phosphate and total polyphenols. Support vector machines were applied to these variables allowing the differentiation of the three considered geographical origins with an overall prediction ability of 99.3%.