近红外光谱结合主成分分析和聚类分析鉴别炉甘石生品、伪品和炮制品

Identification of Crude Products,Counterfeit Products and Processed Products of Calamina by Near Infrared Spectroscopy,Principal Component Analysis and Cluster Analysis

目的：利用主成分判别分析和聚类分析建立炉甘石生品、伪品和炮制品的近红外光谱鉴别模型。方法：采集炉甘石生品、伪品和炮制品的近红外光谱,每类样品随机划分为训练集和测试集。对光谱预处理方法和建模谱段进行筛选,分别建立主成分判别分析模型及聚类分析模型。结果：光谱经一阶导数预处理,主成分判别分析模型的特征谱段为4 800～4 000 cm^-1,聚类模型的特征谱段为7 300～7 000,4 800～4 000 cm^-1。在主成分判别分析模型中,预测准确率94.34%;在聚类模型中,模型的预测准确率96.23%。结论：所建立的近红外主成分判别分析模型和聚类分析模型均可用于炉甘石生品、伪品和炮制品的鉴别。

Objective： To establish a near infrared spectral discriminant model of crude products,counterfeit products and processed products of Calamina by principal component analysis and cluster analysis.Method： Near infrared spectra of crude products,counterfeit products and processed products of Calamina were collected. Each category sample was randomly divided into training set and testing set. The spectral preprocessing methods and modeling spectral bands were screened,the principal component discriminant analysis model and the cluster analysis model were established respectively. Result： Spectrs were preprocessed by the first derivative.The characteristic spectral band of the principal component discriminant analysis model was 4 800-4 000 cm^-1,and the characteristic spectral bands of the cluster analysis model were 7 300-7 000,4 800-4 000 cm^-1. In the principal component discriminant analysis model, the prediction accuracy rate was 94. 34%. In the cluster analysis model,the prediction accuracy rate was 96. 23%. Conclusion： The principal component analysis model and cluster analysis model of near infrared spectra can be used for identification of crude products,counterfeit products and processed products of Calamina.