Chemometrics for Food Authenticity Applications

Verification of food authenticity is a complex issue that generally involves the determination of many physical or chemical properties. The best way to get useful information from such data is to process them by multivariate pattern recognition methods, which allow consideration of intercorrelations among variables-a key issue for an accurate characterization of any system under study. In fact, multivariate pattern recognition approaches are usually required to obtain efficient models. Pattern recognition methods can be used for different purposes but mainly for data exploration and the development of predictive models. The output of these models can be qualitative or quantitative. A proper validation of predictive models is always required to provide reliable predictions.