基于人工智能嗅觉技术和化学计量学的白及饮片真伪鉴别  

Authenticity Identification Method of Bletilla Rhizoma Decoction Pieces Based on Artificial Intelligence Senses Technology and Chemometrics

作　　者：桂新景 李涵 王艳丽[1] 范雪花 李海洋 侯富国 姚静[1] 张璐[1] 施钧瀚[1] 刘瑞新[1,2,3,4,5] 李学林 GUI Xinjing;LI Han;WANG Yanli;FAN Xuehua;LI Haiyang;HOU Fuguo;YAO Jing;ZHANG Lu;SHI Junhan;LIU Ruixin;LI Xuelin(Department of Pharmacy,the First Af liated Hospital of Henan University of Chinese Medicine,Zhengzhou 450000,China;School of Pharmacy,Henan University of Chinese Medicine,Zhengzhou 450046,China;Henan Province Engineering Research Center for Clinical Application,Evaluation and Transformation of Traditional Chinese Medicine,Zhengzhou 450000,China;Co-construction Collaborative Innovation Center for Chinese Medicine and Respiratory Diseases by Henan&Education Ministry of P.R.China,Henan University of Traditional Chinese Medicine,Zhengzhou 450046,China;Henan Provincial Key Laboratory for Clinical Pharmacy of Traditional Chinese Medicine,Zhengzhou 450000,China)

机构地区：[1]河南中医药大学第一附属医院药学部,郑州450000 [2]河南中医药大学药学院,郑州450046 [3]河南省中药临床应用、评价与转化工程研究中心,郑州450000 [4]河南中医药大学呼吸疾病中医药防治省部共建协同创新中心,郑州450046 [5]河南省中药临床药学中医药重点实验室,郑州450000

出　　处：《医药导报》2024年第3期441-451,共11页Herald of Medicine

基　　金：河南省科技攻关项目(222102310377);国家重点研发计划中医药现代化重点专项课题(2017YFC1703400,2017YFC1703401,2017YFC1703402);河南省中医药科学研究专项课题(2023ZY2034);河南省中医药拔尖人才培养项目(重点项目)(2019ZYBJ07);河南省高层次人才特殊支持“中原千人计划”-“中原青年拔尖人才”项目(ZYQR201912158);河南省卫生健康中青年学科带头人专项(HNSWJW-2020014);河南省卫生健康委员会国家中医临床研究基地科研专项课题(2021JDZY104)。

摘　　要：目的采用人工智能嗅觉技术结合化学计量学方法,探索建立适合鉴别白及饮片真伪的新方法。方法收集白及及其易掺伪饮片(天麻、玉竹、黄花白及)134批,以α-FOX4000电子鼻嗅觉感官数据为自变量(X),依据2020年版《中华人民共和国药典》和地方标准鉴别,参考传统经验辨识结果作为标杆辨识信息(Y),采用主成分分析-判别分析(PCA-DA)、偏最小二乘-判别分析(PLS-DA)、支持向量机(SVM)、最小二乘-支持向量机(LS-SVM)以及反向传播神经网络(BP-NN)5种化学计量学方法建立白及与非白及的二分类辨识模型和4种饮片的四分类辨识模型Y=F(X);以鉴别准确率为指标,优选最优分类辨识模型。结果经留一法交互验证,在二分类辨识模型中,PCA-DA、PLS-DA、SVM和BP-NN正确率分别为95.52%、97.01%、91.79%和84.33%,以PLS-DA最优;在四分类辨识模型中,PCA-DA、PLS-DA、LS-SVM和BP-NN正确率分别为91.04%、88.06%、89.55%和82.28%,以PCA-DA最优。结论运用电子鼻技术对白及及其近似饮片进行气味特征的表征,并与多元统计分析方法结合,可准确、快速地鉴别白及饮片,为中药饮片气味客观化表达及真伪鉴别提供了新的思路和方法。Objective To establish a rapid identification method for Bletilla Rhizoma and its similar decoction pieces using artificial intelligence nose technology and chemometric methods.Methods A total of 134 batches of Bletilla Rhizoma and its approximate decoction pieces(Gastrodiae Rhizoma,Polygonati odorati Rhizoma,and Bletilla ochracea)were collected.Theα-FOX4000 electronic nose olfactory sensory data were set as the independent variable X.The decoction pieces were identified according to the Chinese Pharmacopoeia Commission 2020,local standards,and traditional identification experience.And the identification results were set as the benchmark identification information Y.Five chemometrics methods,principal component analysis-discriminant analysis(PCA-DA),partial least squares-discriminant analysis(PLS-DA),support vector machine(SVM),least squares-support vector machine(LS-SVM)and back propagation neural networks(BP-NN)were used to establish the two-class identification models of Bletilla striata and non-Bletilla striata and the four-class identification models of four kinds of decoction pieces Y=F(X).Taking the identification accuracy rate as the index,the optimal classification identification model is optimized.Results The positive discrimination rates of PCA-DA、PLS-DA、SVM and BP-NN in the two classification identification model were 95.52%,97.01%,91.79%and 84.33%,respectively.The PLS-DA identification model showed best accuracy.In the four classification identification model,the positive discrimination rates of PCA-DA、PLS-DA、LS-SVM and BP-NN models were 91.04%,88.06%,89.55%and 82.28%respectively.And PCA-DA identification model was the best.Conclusion The electronic nose technology can be used to characterize the odor characteristics of Bletilla Rhizoma and its approximate decoction pieces,and combined with multivariate statistical analysis methods,it can accurately and quickly identify Bletilla Rhizoma,providing a new idea and method for the objective expression of the odor of Chinese herbal decoction piece

关 键 词：白及 电子鼻 化学计量学 鉴别 

分 类 号：R286[医药卫生—中药学] R927.1[医药卫生—中医学]