基于多元描述性统计对葡萄酒质量的评价

A Study of the Evaluation of Wine Quality Based on the Multiple Descriptive Statistics

基于对两组评酒员的评价结果的方差分析,以及显著性检验,第二组评价结果的可信度较之第一组的更高.在以此确定葡萄酒质量的基础上,本文利用SAS软件以酿酒葡萄的各项理化指标为标准对酿酒葡萄样品进行系统聚类,并参照葡萄酒的质量建立了酿酒葡萄的等级体系,分别将红、白酿酒葡萄划分为五个等级;通过典型性相关分析,构建了酿酒葡萄与葡萄酒的理化指标的关系;依据多元回归分析的方法,确立了葡萄酒质量与酿酒葡萄的理化指标之间、葡萄酒质量与葡萄酒的理化指标之间的线性关系.本文认为,由于用于分析研究的方程中的自变量个数偏少,上述关系仅在一定程度上可作为葡萄酒质量的评价标准,而要完全充分评估葡萄酒的质量还需与感官评价相结合.

Based on the analysis of variance of the results of assessments of the two groups of liquor judges,at first,as well as the test of significance,the paper has drawn a conclusion that the assessment of the second liquor judge is much more reliable than the first one.According to the conclusion,this paper does some deeper researches to confirm the quality of wine.Then,the samples of the wine grapes have been clustered systematically by the SAS software according to the standard of various physical and chemical indicators of the wine grapes.Moreover,by reference to the wine quality,red and white wine grapes have been respectively divided into five grades.Through the canonical correlation analysis,the relationship of the physical and chemical indicators between wine grapes and wine has been established.Finally,the paper has built the linear relationship between the physical and chemical indicators of the quality of the wine and the wine grapes,also the relationship between the indicators of the quality of the wine and the wine.However,because of limited independent variables in the equation,it is only to some not full degree that the above relationships could be applied to evaluate the quality of wine.Thus,in order to evaluate the quality of wines better,it is necessary to combine with the sensory evaluation.