基于电子感官和机器学习模型的混合金银花区分鉴别方法研究  

Research on the Discrimination Method of Hybrid Honeysuckle Based on the Electronic Sensory and Machine Learning Model

作　　者：杨天歌 倪诗婷 高旭华 潘福璐 陶欧[1] YANG Tian-ge;NI Shi-ting;GAO Xu-hua;PAN Fu-lu;TAO Ou(School of Chinese Materia Medica,Beijing University of Chinese Medicine,Beijing 102488,China)

机构地区：[1]北京中医药大学中药学院,北京102488

出　　处：《特产研究》2021年第3期19-22,27,共5页Special Wild Economic Animal and Plant Research

基　　金：北京中医药大学校级横向科研发展基金(2020072220051)。

摘　　要：为了考察电子感官特征结合机器学习模型区分鉴别不同花期混合金银花的效能,将金银花完全开花花朵和未完全开花花蕾以不同比例混合后粉碎,使用扫描仪和电子舌分别获取视觉、味觉感官特征,利用主成分分析、k最近邻、决策树、支持向量机、随机森林和梯度提升树分析方法对数据对比研究,考察区分鉴别效能得出最优模型。结果表明,主成分分析方法不能区分不同花期混合后的金银花;支持向量机模型可以实现对混合金银花的区分鉴别,最高正确率为88%。得知利用支持向量机模型可以实现基于电子感官的不同花期混合金银花的快速鉴别。In order to examine the discrimination performance of mixed honeysuckle at different flowering periods using electronic sensory features combined with machine learning models,the fully-flowered flowers and incompletely-flowered buds of the honeysuckle are mixed in different proportions and then crushed.The scanner and the electronic tongue are used to obtain the visual and taste sensory characteristics respectively.Principal component analysis,k-nearest neighbor,decision tree,support vector machine,random forest and gradient boosting tree analysis methods are used to compare data,and obtain the optimal model which performs the highest discrimination performance.The results show that the principal component analysis method can’t distinguish honeysuckle mixed with different flowering periods;The support vector machine model can distinguish mixed honeysuckle with the highest accuracy rate of 88%.It is known that the use of support vector machine model can realize the rapid identification of mixed honeysuckle in different flowering periods based on the electronic sensory.

关 键 词：电子舌 视觉特征 金银花 支持向量机 

分 类 号：S567.79[农业科学—中草药栽培] R282.5[农业科学—作物学]