Hou Xuewen, Wang Guangli, Wang Xin, Ge Xinmin, Fan Yiren, Jiang Rui, Nie Shengdong
School of Medical Instrument and Food Engineering, University of Shanghai for Science and Technology, Shanghai, China.
School of Geosciences, China University of Petroleum, Qingdao, China.
J Sci Food Agric. 2021 Apr;101(6):2389-2397. doi: 10.1002/jsfa.10862. Epub 2020 Oct 24.
As extra virgin olive oil (EVOO) has high commercial value, it is routinely adulterated with other oils. The present study investigated the feasibility of rapidly identifying adulterated EVOO using low-field nuclear magnetic resonance (LF-NMR) relaxometry and machine learning approaches (decision tree, K-nearest neighbor, linear discriminant analysis, support vector machines and convolutional neural network (CNN)).
LF-NMR spectroscopy effectively distinguished pure EVOO from that which was adulterated with hazelnut oil (HO) and high-oleic sunflower oil (HOSO). The applied CNN algorithm had an accuracy of 89.29%, a precision of 81.25% and a recall of 81.25%, and enabled the rapid (2 min) discrimination of pure EVOO that was adulterated with HO and HOSO in the volumetric ratio range of 10-100%.
LF-NMR coupled with the CNN algorithm is a viable candidate for rapid EVOO authentication. © 2020 Society of Chemical Industry.
由于特级初榨橄榄油（EVOO）具有很高的商业价值，它经常被其他油类掺假。本研究调查了使用低场核磁共振（LF-NMR）弛豫测量法和机器学习方法（决策树、K近邻、线性判别分析、支持向量机和卷积神经网络（CNN））快速识别掺假EVOO的可行性。
LF-NMR光谱有效地将纯EVOO与掺有榛子油（HO）和高油酸向日葵油（HOSO）的EVOO区分开来。应用的CNN算法准确率为89.29%，精确率为81.25%，召回率为81.25%，能够在2分钟内快速鉴别出体积比在10%-100%范围内掺有HO和HOSO的纯EVOO。
LF-NMR与CNN算法相结合是快速鉴定EVOO的可行方法。©2020化学工业协会。
