“…This implies the development of predictive discriminant models on the sensory analysis dataset previously obtained. For example, linear discriminant analysis (LDA) was used to recognize smoked food products (Ojeda, Barcenas, Perez-Elortondo, Albisu, & Guillen, 2002) and artificial neural networks (ANN) have been trained to discriminate between original Scotch whisky and similar products (Jack & Steele, 2002). Over the last decade, the machine learning and statistical communities have developed new modeling tools based on ensembles (or groups) of predictors (see Breiman, 1996;Breiman, 2001;Freund & Schapire, 1996, or reviews from Ho, 2002Tresp, 2001) that are consistently more accurate than single predictors (Bauer & Kohavi, 1999).…”