青砖茶风味轮及滋味类型模型的构建  

Construction of flavor wheel and taste type model for Qingzhuan tea

作　　者：赖建东 马梦君 余子铭 程繁杨 郭志明 邱首哲 LAI Jiandong;MA Mengjun;YU Ziming;CHENG Fanyang;GUO Zhiming;QIU Shouzhe(South Hubei Culture Research Center,Hubei University of Science and Technology,Xianning 437100,Hubei,China;Xianning Academy of Agricultural Sciences,Xianning 437100,Hubei,China)

机构地区：[1]湖北科技学院鄂南文化研究中心,湖北咸宁437100 [2]咸宁市农业科学院,湖北咸宁437100

出　　处：《粮食与油脂》2025年第4期91-96,125,共7页Cereals & Oils

基　　金：国家现代农业产业技术体系(CARS-19);咸宁市科技人才服务专项(2024RCFW011);湖北科技学院校企合作横向课题(2023HX209);湖北科技学院2024年度“双百工程”(36)。

摘　　要：为客观准确地评价青砖茶的品质,对47个青砖茶进行了感官评价和品质成分检测,构建了青砖茶风味轮和滋味类型贝叶斯(Bayes)预测模型。结果表明:感官评价可将青砖茶分为陈香醇正型、药香醇厚型2个类型;同时以品质成分数据为基础,进行主成分分析(PCA)和二级聚类分析,结果与感官评价分类结果一致;滋味类型Bayes预测模型自身验证正确率为100%,交叉验证正确率为94%,回判检测正确率为83.33%,对青砖茶滋味类型分类具有较好的预测判别效果,可为青砖茶滋味类型量化提供参考。In order to objectively and accurately evaluate the quality of Qingzhuan tea,the sensory evaluation and quality components of detection 47 samples of Qingzhuan tea were conducted.A flavor wheel and a Bayes prediction model for the taste types of Qingzhuan tea were developed.The results showed that sensory evaluation could classify Qingzhuan tea into two distinct types as the“Chenxiang Chunzheng”type and the“Yaoxiang Chunhou”type.In parallel,principal component analysis(PCA)and hierarchical clustering analysis were carried out based on the data of quality components,and the results aligned with the sensory evaluation classification.The Bayes prediction model for taste types exhibited 100%accuracy rate in self-validation,94%accuracy rate in cross-validation,and 83.33%accuracy rate in backtesting.The Bayes prediction model had a good prediction and discrimination effect on the classification of taste types of Qingzhuan tea,and could provide reference for the quantification of taste types of Qingzhuan tea.

关 键 词：青砖茶 感官评价 风味轮 判别模型 

分 类 号：TS272.54[农业科学—茶叶生产加工]