Based on 551 Raman spectral data of 6 categories (38 brands) of vegetable oils, orthogonal partial least squares discriminant analysis (OPLS-DA), and support vector machine (SVM) models were developed to compare the effects of successive projections algorithm (SPA) and competitive adaptive reweighted sampling (CARS).The accuracy of the improved OPLS-DA model based on two feature extraction algorithms was 82.53% and 83.13%, respectively, which was lower than the model based on full spectrum and SVM models.Brand classification of vegetable oils was investigated using CARS combined with SVM.For the test set, the accuracy for corn, olive, sunflower, and sesame oils was 100%, while the accuracy for coconut oil and peanut oil brands was 22.22% and 63.64%, respectively.Both algorithms could reduce the number of variables and computational resources required to build the classification model.Replacing full-spectrum data with lower spectral data may lead to a decrease in recognition accuracy.The SVM outperformed OPLS-DA in solving multiple classification problems with high sample similarity.The difference in accuracy between vegetable oils may be related to the production process and vegetable oil ingredients used by the manufacturer.The SVM model based on Raman spectroscopy and feature extraction algorithm provides a reference for the nondestructive and rapid inspection of vegetable oils.
SU Dongbin
,
QIN Jiahui
,
LI Kaikai
. Raman spectroscopy combined with machine learning for classification of vegetable oils[J]. Food and Fermentation Industries, 2024
, 50(6)
: 274
-281
.
DOI: 10.13995/j.cnki.11-1802/ts.036153
[1] 卢晓蕊, 班楠, 崔杉, 等. 食用植物油抽检过程质量控制措施[J]. 食品安全质量检测学报, 2021, 12(19):7813-7817.
LU X R, BAN N, CUI S, et al. Quality control measures for sampling process of edible vegetable oil[J]. Journal of Food Safety & Quality, 2021, 12(19):7813-7817.
[2] PAN F G, YANG E Q, CHEN X M, et al. Identification of adulterated evening primrose oil based on GC-MS and FT-IR combined with chemometrics[J]. European Journal of Lipid Science and Technology, 2022, 124(10): 2200066.
[3] TROYA F, LERMA-GARCÍA M J, HERRERO-MARTÍNEZ J M, et al. Classification of vegetable oils according to their botanical origin using n-alkane profiles established by GC-MS[J]. Food Chemistry, 2015, 167:36-39.
[4] ZHAO X, MA F, LI P W, et al. Simultaneous determination of isoflavones and resveratrols for adulteration detection of soybean and peanut oils by mixed-mode SPE LC-MS/MS[J]. Food Chemistry, 2015, 176:465-471.
[5] CRIADO-NAVARRO I, MENA-BRAVO A, CALDERÓN-SANTIAGO M, et al. Determination of glycerophospholipids in vegetable edible oils: Proof of concept to discriminate olive oil categories[J]. Food Chemistry, 2019, 299:125136.
[6] VYVIURSKA O, JÁNOŠKOVÁ N, JAKUBÍK T, et al. Comprehensive two-dimensional gas chromatography-mass spectrometry analysis of different types of vegetable oils[J]. Journal of the American Oil Chemists’ Society, 2015, 92(6):783-790.
[7] AGHILI N S, RASEKH M, KARAMI H, et al. Detection of fraud in sesame oil with the help of artificial intelligence combined with chemometrics methods and chemical compounds characterization by gas chromatography-mass spectrometry[J]. LWT, 2022, 167:113863.
[8] TRUZZI E, MARCHETTI L, BENVENUTI S, et al. Novel strategy for the recognition of adulterant vegetable oils in essential oils commonly used in food industries by applying 13C NMR spectroscopy[J]. Journal of Agricultural and Food Chemistry, 2021, 69(29):8276-8286.
[9] ANCORA D, MILAVEC J, GRADIŠEK A, et al. Sensitivity of proton NMR relaxation and proton NMR diffusion measurements to olive oil adulterations with vegetable oils[J]. Journal of Agricultural and Food Chemistry, 2021, 69(41):12081-12088.
[10] KHARBACH M, KAMAL R, MARMOUZI I, et al. Fatty-acid profiling vs UV-Visible fingerprints for geographical classification of Moroccan Argan oils[J]. Food Control, 2019, 95:95-105.
[11] SHIMAMOTO G G, TUBINO M. Alternative methods to quantify biodiesel in standard diesel-biodiesel blends and samples adulterated with vegetable oil through UV-Visible spectroscopy[J]. Fuel, 2016, 186:199-203.
[12] WU X J, GAO S B, NIU Y D, et al. Quantitative analysis of blended corn-olive oil based on Raman spectroscopy and one-dimensional convolutional neural network[J]. Food Chemistry, 2022, 385:132655.
[13] HUANG F R, LI Y P, GUO H X, et al. Identification of waste cooking oil and vegetable oil via Raman spectroscopy[J]. Journal of Raman Spectroscopy, 2016, 47(7):860-864.
[14] JOHNSON J B, THANI P R, MANI J S, et al. Mid-infrared spectroscopy for the rapid quantification of eucalyptus oil adulteration in Australian tea tree oil (Melaleuca alternifolia)[J]. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2022, 283:121766.
[15] VARGAS JENTZSCH P, CIOBOTĂ V. Raman spectroscopy as an analytical tool for analysis of vegetable and essential oils[J]. Flavour and Fragrance Journal, 2014, 29(5):287-295.
[16] 黄平捷, 李宇涵, 俞巧君, 等. 基于SPA和多分类SVM的紫外-可见光光谱饮用水有机污染物判别方法研究[J]. 光谱学与光谱分析, 2020, 40(7):2267-2272.
HUANG P J, LI Y H, YU Q J, et al. Classification of organic contaminants in water distribution systems developed by SPA and multi-classification SVM using UV-vis spectroscopy[J]. Spectroscopy and Spectral Analysis, 2020, 40(7):2267-2272.
[17] 聂黎行, 戴忠, 马双成, 等. 近红外光谱技术结合竞争自适应重加权采样算法用于中药定量分析[J]. 中国实验方剂学杂志, 2017, 23(11):45-49.
NIE L X, DAI Z, MA S C, et al. Quantitative analysis of traditional Chinese medicine by near-infrared spectroscopy combined with competitive adaptive reweighted sampling[J]. Chinese Journal of Experimental Traditional Medical Formulae, 2017, 23(11):45-49.
[18] ARAÚJO M C U, SALDANHA T C B, GALVÃO R K H, et al. The successive projections algorithm for variable selection in spectroscopic multicomponent analysis[J]. Chemometrics and Intelligent Laboratory Systems, 2001, 57(2):65-73.
[19] 赵昱萱, 黄威, 董林沛, 等. 基于高光谱和化学计量学方法的除草剂残留分类识别研究[J]. 化学研究与应用, 2022, 34(1):91-102.
ZHAO Y X, HUANG W, DONG L P, et al. Identification of herbicide residues based on hyperspectral imaging and chemometrics[J]. Chemical Research and Application, 2022, 34(1):91-102.
