A commercial electronic nose and an home made electronic tongue were used, together with spectrophotometric determination of phenolic compounds and color, in order to predict the sensorial descriptors and the overall quality of Italian red dry wines of different denominations of origin. An expert wine tester having an O.N.A.V. certificate, selected the wines on the basis of their well-known sensorial characteristics, e.g. astringency, bitterness, aroma, color, body, etc. The analytical data were used to build predictive models of the sensorial descriptors by means of the Genetic Algorithms, proposed as an alternative to the mostly used PLS analysis and employed to select subsets of variables that maximize the predictive power of regression models. On the selected models accurate validation techniques such as Leave-one-out and Bootstrap were applied. The present work demonstrate the possibility of using innovative devices as the electronic nose and the electronic tongue to obtain the sensorial properties of wines. Moreover, the genetic algorithms could represent a rational operative procedure for building regression models with real predictive capability.
Tecniche innovative combinate “naso elettronico” e “lingua elettronica” per la predizione di descrittori sensoriali di vini rossi secchi mediante l’uso degli Algoritmi Genetici / S. Buratti, S. Benedetti, D. Ballabio. - In: INGREDIENTI ALIMENTARI. - ISSN 1594-0543. - 5(2006 Nov), pp. 6-9.
Tecniche innovative combinate “naso elettronico” e “lingua elettronica” per la predizione di descrittori sensoriali di vini rossi secchi mediante l’uso degli Algoritmi Genetici
S. BurattiPrimo
;S. BenedettiSecondo
;
2006
Abstract
A commercial electronic nose and an home made electronic tongue were used, together with spectrophotometric determination of phenolic compounds and color, in order to predict the sensorial descriptors and the overall quality of Italian red dry wines of different denominations of origin. An expert wine tester having an O.N.A.V. certificate, selected the wines on the basis of their well-known sensorial characteristics, e.g. astringency, bitterness, aroma, color, body, etc. The analytical data were used to build predictive models of the sensorial descriptors by means of the Genetic Algorithms, proposed as an alternative to the mostly used PLS analysis and employed to select subsets of variables that maximize the predictive power of regression models. On the selected models accurate validation techniques such as Leave-one-out and Bootstrap were applied. The present work demonstrate the possibility of using innovative devices as the electronic nose and the electronic tongue to obtain the sensorial properties of wines. Moreover, the genetic algorithms could represent a rational operative procedure for building regression models with real predictive capability.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.