应用多元回归分析进行葡萄酒质量评价建模

本文针对2012全国大学生数学建模竞赛A题，分析了葡萄酒质量人工品尝存在的不足，进而对若干种葡萄酒的原始数据进行异常值和标准化处理，利用K—W检验、信度分析等方法分析得出红葡萄酒第二组和白葡萄酒第一组的质量评价更为可信；通过相关性分析筛选芳香物质的成分，运用 TOPSIS方法的秩和排序对各类葡萄酒质量划分等级，对理化指标两独立样本进行Mann－Whitney U 检验；最后建立多元回归模型分析得出评价酿酒葡萄和葡萄酒的理化指标对葡萄酒质量的影响关系，并通过Wilcoxon符号秩检验其模型的可行性。

Wine Quality Evaluation Model Based on Multivariate Regression Analysis

Aimed at the 2012 China Undergraduate Mathematical Contest in Modeling Problem A, in this paper, we analyze the wine quality artificial taste of deficiencies and the original data of several kinds of wine enter to outliers and standardization . Firstly, using the K-W test, reliability analysis and so on, it is concluded that the second group and red wine white wine quality evaluation more credible in the first group.Secondly, by using the correlational analyses, the elements of aromatic substances are filtered, and dividing grade is carried out according to the method of TOPSIS rank and sorting the quality of several kinds of wine. Then, the physical and chemical indexes make the Mann-Whitney U two independent sample tests.Finally, we establish a Multi-variate Linear Regression Model, to explore the relationship of the quality evaluation between the grape and wine＆#39;s indexes and the quality of the wine.It is available that the Multivariate Linear Regression Model is tested according to the Wilcoxon signed rank test.