基于最小二乘支持向量机的橄榄油掺杂拉曼快速鉴别方法

提出了一种基于最小二乘支持向量机(LS-SVM)的橄榄油掺杂拉曼快速鉴别方法。首先,收集若干己知类别的橄榄油样作为训练样本,获取其拉曼谱图,并对其谱图进行预处理和波段选择,进而构建LSSVM分类器;对于未知类别的油样,获取其拉曼谱图,并进行相应的预处理和波段选择,由LSSVM分类器获得鉴别结果。实验以7种已知的特级初榨橄榄油为基础,分别掺入4种其它植物油(大豆油、菜籽油、玉米油、葵花籽油),获得112个掺杂油样。将全部样本随机分成训练集和测试集,对测试集样本的预测实验结果表明,本文方法能有效鉴别橄榄油掺杂,且掺杂量最低检测限为5%。与其它分类方法相比,LSSVM分类法具有最佳的分类性能。该方法快速、简便,为橄榄油掺杂鉴别提供了一种全新的方法。

Fast Discrimination of Olive Oil Adulteration Based on Raman Spectra Using Least Squares Support Vector Machine

A fast discrimination method to olive oil adulteration based on Raman spectra using least squares support vector machine LSSVM was presented.Firstly,some known class olive oil samples were chosen randomly as training samples and their original Raman spectra were obtained,then a pretreatment and band selection were made for those spectra,and then,the LSSVM classifier was built.Secondly,for the Raman spectra of unknown test samples,the same pretreatment and band selection were used.Finally,the discrimination results were attained through the LSSVM classifier.The experiment was based on seven known Extra virgin olive oil and 112adulterated samples were acquired by mixing four other vegetable oils(soybean,rapeseed,corn and sunflower oil)into the basic oils.The whole samples were divided into training test and testing test randomly,the test result shows that this method was able to discriminate olive oil adulteration and the lowest detection limit of the doping amount was 5%.Compared with other classification methods,LSSVM classifier has the best classification performance.The above method provided a new solution for discriminate olive oil adulteration,which was fast and pretty easy.