神经网络技术在化橘红两品种识别中的应用  

Application of Neural Network in the Discrimination of Two Different Species of Exocarpium Citrus Grandis

作　　者：魏航[1] 林励[2] 黄志煜[1] 陈永刚[2] 袁旭江[2] 

机构地区：[1]广州中医药大学信息技术学院,广东广州510405 [2]广州中医药大学中药学院,广东广州510405

出　　处：《广州中医药大学学报》2011年第3期272-276,共5页Journal of Guangzhou University of Traditional Chinese Medicine

基　　金：国家科技部"十二五"科技支撑计划项目(编号:2011BAIOIB02)

摘　　要：【目的】建立有效区别毛橘红与光橘红2种药材的识别模型。【方法】收集不同产地的23批化橘红药材样品的指纹图谱,采用主成分分析法提取主成分,利用BP神经网络进行模式识别。【结果】建立了有效识别毛橘红和光橘红的神经网络模型,有效识别率超过91.3%,其中毛橘红均能被正确识别。【结论】神经网络技术可有效识别出道地药材毛橘红。Objective To establish an effective model for the discrimination of Citrus grandis ‘Tomentosa’ and C.grandis（L.） Osbeck,which are the medicinal plant sources of Exocarpium Citrus Grandis.Methods Twenty-three batches of medicinal material samples of Exocarpium Citrus Grandis were collected from different places.Fingerprints of the samples were determined by HPLC,and several variables based on principal component analysis were selected for the establishment of BP neural network model.Results An effective BP neural network model was established.The distinguishing rate was over 91.3%,and the samples of Citrus grandis‘Tomentosa’ were all correctly recognized.Conclusion Neural network technology can offer a feasible approach to the effective discrimination of Citrus grandis ‘Tomentosa’

关 键 词：化橘红/化学 柚皮苷/分析 野漆树苷/分析 BP神经网络 色谱法 高压液相 指纹图谱 

分 类 号：R282.5[医药卫生—中药学]