一种基于化学计量学方法的大曲质量分类鉴别方法

A research of Identifying Daqu's Quality Based on Chemometric Methods

大曲质量的科学、准确判定是白酒品质分析的重点和难点问题之一。实现大曲质量的科学、准确鉴定,对促进白酒健康产业的快速发展具有重要的意义。研究以选取四房出房曲的水分、酸度、糖化力、感官评价得分组成的原始数据矩阵为基础,先使用DRAUG算法计算数据矩阵的主因子数,再采用奇异值分解算法,结合方差贡献率和主成分得分计算出各曲块的综合得分,最后采用K-means算法对计算结果进行分类,发现综合得分处于高分段、中分段、低分段的样本,分属于K-means算法分类的类别1、类别2、类别3,证明了两者对各房酒曲的分类结果一致,最终提供了一种科学、准确的基于化学计量学方法的大曲质量分类鉴别方法。

It is an important and difficult research point to recognize Daqu's quality scientifically and accurately in the field of liquor analyzing,which is also of great significance to the rapid development of the liquor and health industry.The research is based on the original data matrix including moisture,acidity,saccharifying power and sensory evaluation's marks.Firstly DRAUG algorithm is applied to calculate the number of primary factors,then they figure out Daqu's composite scores by SVD algorithm,variance proportion and principal component scores,finally Daqu is classified by adopting K-means algorithm.Samples whose composite scores are in the field of high section,middle section and low section belong to Category 1,Category 2,Category 3 in K-means algorithm respectively.The two categorization approach's calculation results are consistent.This paper provides a scientifical and accurate way which recognizes Daqu's quality by chemometric methods.