期刊
FOOD CONTROL
卷 17, 期 3, 页码 197-206出版社
ELSEVIER SCI LTD
DOI: 10.1016/j.foodcont.2004.10.016
关键词
principal component analysis; Artificial Neural Network; Genetic Algorithm; cluster analysis; Ossolano cheese
yOssolano is a typical Italian cheese produced in the Piedmont region. The authenticity of this cheese variety is becoming more important with the request of the PDO label (Protected Designation of Origin). The aim of this paper is to present a method for an efficient and accurate prediction of the Ossolano cheese production origin. A further objective of this work is to reduce the number of analyses needed to cluster Ossolano samples, so that chemometrics methods like PCA and Artificial Neural Networks (ANNs) can be applied routinely to identify this cheese. The Neural Network presented in this paper is excellent in predicting the origin of the Ossolano cheese. Chemical parameters such as the composition of fatty acids and proteolytic profiles obtained by electrophoresis and chromatography, as well as biochemical parameters, allowed us to distinguish the samples with a high success rate. Instead biogenic amines and their precursor amino acids were not enough to distinguish the samples because the corresponding success rate was low. In order to identify the most useful variables and to cut parameters irrelevant for classification purposes, we used a Genetic Algorithm (GA). Using only half of the variables, without making any a priori assumption about their importance, we were able to achieve a success rate comparable to the one obtained with all the variables. (c) 2005 Elsevier Ltd. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据