基于神经网络权值抽取与粒度一致性的葡萄酒的质量评价探讨

摘要

能否用葡萄和葡萄酒的理化指标来评价葡萄酒的质量，这是近年来葡萄酒评价领域的热门研究课题。为了研究这一问题，本文先用主成份分析简化理化指标和芳香物质，建立BP神经网络确定这些指标关于葡萄酒质量的权重。在分析理化指标是否能评价葡萄酒质量的问题上，我们引入粒度一致性的概念，结合线性加权分析法与灰度关联分析，对评价问题进行粒度一致性分析。实验数据证明，葡萄和葡萄酒的理化指标对葡萄酒质量有影响，但不能全面评价葡萄酒质量。本文所使用的该种组合分析方法，对葡萄酒质量的量化评价具有重要指导意义。%Discussing the possibility of using physical and chemical indexes to evaluate wine quality is the popular research topic in the field of wine evaluation recently .To further study this topic ,princi‐pal components are proposed in this paper in the analysis of aromatic substance and simplified physical and chemical indexes ,then BR neural network is established to determine the index weight of wine qual‐ity .When discussing whether physical and chemical indexes can be used to evaluate wine quality ,we in‐troduce the concept ,consistency of particle size .Along with linear weighted analysis and gray relation analysis ,the evaluation process is analyzed based on its consistency of particle size .Experimental statis‐tics show that physical and chemical indexes of grapes and wine have effects on wine quality but can not give a full appraisal of it .The analysis methods applied in this paper have important guiding significance on the study of the quantitative evaluation of wine quality .