基于主成分分析和判别分析的咖啡掺假鉴别方法研究

Study of the Identification Method of Coffee Adulteration Based on Principal Component Analysis and Discriminant Analysis

为建立基于主成分分析和判别分析的咖啡掺假鉴别方法,利用离子色谱/脉冲安培检测法测定烘焙咖啡及咖啡掺假物中糖类标记物含量,采用SPSS统计软件进行主成分分析和判别分析,建立鉴别掺假物种类的方法。结果表明:主成分分析中,样品前2个主成分的累计贡献率达到90.055%,已含样本的大部分信息。前2个主成分构成的二维得分图显示,15个模拟掺假样品按掺假种类分成3组,从而可对掺假样品进行分类识别。通过判别分析,咖啡掺黑玉米、咖啡掺大麦在掺假物含量为5%时已能被正确判别分组,咖啡掺莓果粉在掺假物含量为10%时已能被正确判别分组,准确率达到93.3%。该方法准确性良好、操作简便,可作为一种咖啡掺假鉴别评价方法。

In order to establish the identification method of coffee adulteration based on principal component analysis and discriminant analysis,the content of sugar markers in roasted coffee samples and coffee adulterants was determined by ion chromatography/pulsed amperometric detection.The principal component analysis and discriminant analysis methods were carried out by SPSS statistical software,and the method of identifying the types of adulterants was established.The results show that in principal component analysis,the cumulative contribution rate of the first two principal components of the sample reaches 90.055%,which contains most of the information of the sample.The two-dimensional score diagram composed of the first two principal components shows that 15 simulated participating samples are divided into three groups according to the adulteration type,so the adulterated samples can be classified and identified.Through discriminant analysis,coffee mixed with black corn and coffee mixed with barley can be correctly classified and grouped when the adulterant content is 5%,and coffee mixed with berry powder can be correctly classified and grouped when the adulterant content is 10%,with an accuracy of 93.3%.The method has good accuracy and simple operation.It can be used as a method for the identification and evaluation of coffee adulteration.