葡萄酒评价模型的设计与求解.docVIP

葡萄酒评价模型的设计与求解 黄亚坤, 韩磊，王梦瑶 （安徽师范大学数学计算机科学学院，安徽，芜湖，241000） 关键词：T检验；改进K均匀聚类分析；模拟退火；广义回归神经网络 摘 要：针对葡萄酒质量评价的多样性和复杂性等问题，围绕评酒员评价的差异性、酿酒葡萄的分级、理化指标对葡萄酒质量的影响分别建立模型，并对结果进行了详细的分析。首先采用置信区间法降低同一酒样的变异系数，按照其方差贡献率进行综合评分，在此基础上，通过组内评价指标变异度的检验，进行二次方差检验两组评分结果的可信度；对于葡萄酒的分类，利用数据挖掘提取方法，得出主成分，为了改进K均值聚类算法的局限性，提高聚类的有效性；最后考虑多维变量之间的关系，提出了基于广义神经网络模型，研究酿酒葡萄和葡萄酒的理化指标对葡萄酒质量的影响程度。并通过实际数据进行仿真，结果显示了提出的模型具有一定的合理性和有效性。 中图分类号：O235 文献标识码：A 文章编号： (2013) 04 The design and solution of wine evaluation model Huang Yakun, Han Lei, Wang Yang. (School ofMathematics Computer Science, Anhui Normal University, Wuhu241000,China ) Key words：T inspect; Improved K uniform clustering analysis; Simulated annealing; Generalized regression neural network Abstract：In this paper , we aim at the issue of quality assessment , and center on the Significant difference of the tasting members’ evaluation results and reliability、classification of wine grape、the contact between the physical and chemical indicators of the wine grape and wine、the effect and evaluation of the physical and chemical indicators of the wine grape and wine to the quality of the port wine , build models respectively and do a detailed analysis of the result. Firstly we using the confidence interval method to reduce the coefficient of variation of the same wine sample, in accordance with its variance contribution rate of the composite score, on this basis, we do the second variance test two sets of ratings results credible by group evaluation variability inspection; To the classification of wine, our data mining extraction method obtained the principal components, in order to improve the limitations of the K-means clustering algorithm, to improve the effectiveness of the clustering simulated annealing; Finally, we consider the relationship between the multi-dimensional variables, we propose training relevant sample data based on generalized neural network model to study the impact of the physical and chemical indicators of the quality of t