“…With the development of modern instruments more and more techniques have been used in the metabolomics study including chromatography, mass spectrometry, nuclear magnetic resonance (NMR) (Cevallos-Cevallos, Reyes-De-Corcuera, Etxeberria, Danyluk, & Rodrick, 2009). Among these methods, NMR-based metabolomics has proved a promising tool for food compositional analysis and food material characterization due to its capability to determine a great number of metabolites rapidly and simultaneously (Hill, Dias, & Roessner, 2016;Soininen et al, 2014). Currently, many multivariate data analyses approaches used in combination with 1 H-NMR metabolomics have been indicated as versatile analytical tools that can provide comprehensive information on the metabolic profiles in both vegetal and animal matrices, such as principal component analysis (PCA) (Francini et al, 2017), partial least-squares discriminant analysis (PLS-DA) (Pereira et al, 2005), Linear Discriminant Analysis (LDA) (Huo et al, 2017), k-Nearest Neighbor (KNN) (Xu, Song, Li, & Wan, 2012), Support vector machine (SVM) (Masoum et al, 2007), Least square support vector machine (LS-SVM) (Jin et al, 2015).…”