藜麦产地的红外光谱鉴别

种植环境差异导致不同产地的藜麦有差异,故对不同产地的藜麦进行区分鉴别对商家、消费者具有重要参考价值。将中红外光谱与主成分分析(PCA)、线性判别分析(LDA)及混淆矩阵结合对不同产地藜麦进行鉴别研究。结果显示:藜麦的红外光谱主要由淀粉、蛋白质和脂质谱峰组成,且在蛋白质和糖类谱峰上有差异。用600~4000 cm^-1范围的原始光谱进行PCA分析,前两个主成分(PC)取得了92%的累计方差贡献率,基于PCA分析生成的PC进行LDA分析,取得了96.25%的分类精度。基于预测结果的混淆矩阵作为综合评价指标,得到PCA-LDA分类模型的精确度、召回率及特异性分别为96.25%、96.59%和99.48%,说明使用PCA-LDA模型可以对藜麦产地进行有效鉴别。研究表明红外光谱结合多元统计分析方法是鉴别藜麦产地的有效方法。

Identification of Quinoa Origin by Infrared Spectroscopy

Since quinoa differs in speciality from one place to another due to the different planting environment,the identification of quinoa from different producing areas is of great reference value for merchants and consumers.In this paper,infrared spectroscopy combined with principal component analysis(PCA),linear discriminant analysis(LDA)and confusion matrix were used to identify quinoa from different areas.The results showed that the infrared spectra of quinoa were mainly composed of vibrations bands attributed to starch,protein and lipid with differences in bands due to the protein and starch.PCA analysis was carried out on the basis of the original spectra in the range of 600~4000 cm^-1,and 92%cumulative variance contribution rate was obtained on the first two PCs.LDA analysis based on PCs generated from PCA analysis achieved 96.25%classification accuracy.Based on the confusion matrix of the predicted results as a comprehensive evaluation index,the average accuracy,recall rate and specificity of the PCA-LDA classification model are 96.25%,96.59%and 99.48%,respectively,which suggests that the PCA-LDA model could effectively discriminate the origin of quinoa.The present investigation confirms infrared spectroscopy in combination with multivariate statistical analysis to be an effective method to identify the geographical origin of quinoa.