摘要
以低分子有机酸、酯作为条件属性,白酒香型作为决策属性建立了一个知识评判系统,利用粗糙集理论方法讨论了低分子有机酸、酯在白酒各类香型中的重要性,并采用基于属性重要性的约简算法对系统进行了属性简化。结合信息覆盖率最大的原则,选择了一组最佳的简化方案,通过规则约简从而得到了白酒香型的最小决策算法,确定了低分子有机酸酯的含量与各类白酒香型的关系,只需提供较少的属性值就能判别出白酒的香型。
In this study, a knowledge evaluation system had been established by taking low molecular organic acid and ester as the condition at- tributes and liquor flavor types as the decision attributes. The importance of low molecular organic acids in liquor of all kinds of flavor types was discussed by employing rough set theory, and the attributes of the system were simplified by utilizing the reduction algorithm on the basis of the fundamentality of the condition attributes. Then, an optimal simplified program was selected in consideration of the principle of the maximum amount of information coverage. Subsequently, the minimum decision algorithm of Baijiu （liquor） flavor types was obtained by rule reduction. And the relations between the content of low molecular organic acid and ester and Baijiu（liquor） flavor types was confirmed, which meant Baijiu （liquor） flavor type could be identified through only a few provided attribute values.
出处
《酿酒科技》
北大核心
2014年第4期1-4,共4页
Liquor-Making Science & Technology
基金
酿酒生物科技及应用四川省重点实验室基金项目:基于模式识别的年份酒鉴别系统(NJ2010-09)
四川理工学院硕士研究生创新基金项目:白酒识别与质量评估系统(Y201201)
关键词
识别方法
白酒
有机酸酯
白酒香型
信息覆盖率最大
粗集理论
identification
Baijiu（liquor）
organic acid ＆ ester
Baijiu types
maximum amount of information coverage
rough set theory
