摘要
目的:通过核磁共振波谱(NMR)识别灵芝产地。方法:利用正交偏最小二乘判别分析(OPLS-DA)对四个不同产地128个灵芝样品的^(1)H-NMR核磁数据进行分析,并导入支持向量机(SVM)模型进行识别计算。结果:不同产地灵芝按地理来源全部正确分类,模型主成分累计贡献率R^(2)X高达0.968,Q^(2)值达到了0.917,表明所构模型区分能力和预测能力较好,基于核磁数据分类方法可行。进一步将灵芝核磁积分数据导入支持向量机(SVM)进行模式识别,测试集数据100%正确识别,输出数字结果。结论:核磁共振技术与支持向量机联用提供了一种灵芝产地识别的快捷方法,该方法可为中药材的品质分析提供借鉴。
Objective:To recognize the origin of Ganoderma lucidum by nuclear magnetic resonance spectroscopy(NMR).Methods:The ^(1)H-NMR data of 128 samples of Ganoderma lucidum from four different places were analyzed by orthogonal partial least squares discriminant analysis(OPLS-DA)and imported into the support vector machine(SVM)model for identivication calculation.Results:Ganoderma lucidum from different producing areas were correctly classified according to geographical origin,and the cumulative contribution rate R^(2)X of principal components of the model was as high as 0.968,and Q^(2) value was as high as 0.917,indicating that the constructed model had good discriminative and predictive ability,and the classification method based on nuclear magnetic data was feasible.Further,the nuclear magnetic integration data of Ganoderma lucidum was imported into support vector machine(SVM)for pattern recognition.The test set data was 100%correctly recognized and the digital results were output.Conclusion:NMR combined with support vector machine(SVM)provide a fast method to identify the origin of Ganoderma lucidum.This method can provide reference for quality analysis of Chinese medicinal materials.
作者
王寿峰
梁俊威
雍登金
丁秀国
林嘉荣
程纯儒
WANG Shou-feng;LIANG Jun-wei;YONG Deng-jin;DING Xiu-guo;LIN Jia-rong;CHENG Chun-ru(Sichuan University of Science&Engineering,Zigong 643002,China;Sichuan University of Science&Engineering/Liquor Making Biological Technology and Application of Key Laboratory of Sichuan Province,Yibin 644002,China)
出处
《中药材》
CAS
北大核心
2023年第3期598-602,共5页
Journal of Chinese Medicinal Materials
基金
四川省科技厅应用基础研究(2020YJ0404)
酿酒生物技术及应用四川省重点实验室开放基金项目(NJ2018-03)
关键词
灵芝
核磁共振
产地识别
支持向量机
Ganoderma lucidum(Leyss.ex Fr.)Karst.
NMR
Origin identification
SVM
