基于红外光谱概率神经网络的诃子产地鉴别  

Identifying the Origin of Terminalia chebula Retz. Based on Infrared Spectrum and Probability Neural Networks

作　　者：安生梅[1] 吴启勋[1] 吉守祥[1] 

机构地区：[1]青海民族大学化学与生命科学学院,西宁810007

出　　处：《湖北农业科学》2014年第20期4977-4979,共3页Hubei Agricultural Sciences

基　　金：国家自然科学基金项目(81160554);国家教育部春晖计划项目(Z2012108)

摘　　要：利用傅里叶变换红外光谱,测定了不同产地诃子(Terminalia chebula Retz.)样品的红外光谱图。采用常规预处理方法和小波变换对红外光谱原始数据进行了预处理,并采用主成分分析进一步压缩光谱数据,前3个主成分的累积贡献率为98.054%。以前3个主成分作为网络输入,诃子产地类别作为网络输出,建立了概率神经网络,同时对建立该网络模型的扩展常数进行了分析。模型分析表明,建立的网络模型能够对40个诃子样品进行产地鉴别,红外光谱法结合神经网络可作为中药材产地分类鉴别的一种新的现代化方法。Infrared spectrum of Terminalia chebula Retz.from different fields were determined by Fourier transform infrared spectroscopy （FTIR）.The original data matrix of FTIR were pretreated with common preprocessing and wavelet transform.The spectra variables were compressed through the wavelet transformation.The principal component analysis （PCA） method was used to compress the spectral data.The PCA results showed that the first three principal components had the cumulative reliability of 98.054％.The first 3 principal components were used for input nods of probability neural network model and the area sorts of Terminalia chebula Retz.were used for parameters of export.The spread of the probability neural networks model were studied in detail.The model distinguished the producing area of the 40 samples of Terminalia chebula Retz.correctly.The infrared spectral technology combined with the artificial neural networks was proved to be a reliable and practical method for identifying of geographical origin of Lycium barbarum L.It will provide a new reference for the identification of traditional chinese medicine.

关 键 词：诃子(Terminalia chebula Retz.) 傅里叶变换红外光谱 小波变换 概率神经网络 

分 类 号：O657.33[理学—分析化学]