摘要
提出了一种新的葡萄酒质量分级方法。从葡萄和葡萄酒的理化指标、芳香物质出发,首先用逐步回归分析筛选出主要指标;然后用主成分回归分析得到各指标的权重系数;最后引入加权欧氏距离来改进聚类分析,并对葡萄酒质量进行分级。结果表明:该方法将红白葡萄酒均分为三类,分级结果合理,方法具有普适性。
A new method of classification of wine quality was proposed. Based on aromatic matter, physical and chemical criterion of wine and grapes, firstly, the study used the stepwise regression to select main indexes. Then obtained the weight of each index by the principal component regression. Finally, improved the cluster analysis by using weighted Euclidean distance and classed the wine quality. The results show that red wine and white wine are classified into three groups respectively by the above method. The results are reasonable and this method is universalistic.
出处
《佛山科学技术学院学报(自然科学版)》
CAS
2015年第5期17-20,共4页
Journal of Foshan University(Natural Science Edition)
基金
国家自然科学基金资助项目(11271141)
广东省高等教育教学改革项目(GDJG20141038)
关键词
葡萄酒质量分级
逐步回归
主成分回归
聚类分析
权重
classification of wine quality
stepwise regression
principal component regression
cluster analysis
weight