Low-field nuclear magnetic resonance (LF-NMR) and electronic nose combined with Gas chromatography mass spectrometry (GC-MS) were used to collect the data of moisture state and volatile substances to predict the flavor change of ginger during drying. An back propagation artificial neural network (BP-ANN) model was established with the input values of LF-NMR parameters and the output values of sensors for different flavor substances obtained from electronic nose. The results showed that fresh ginger contained three water components: bound water (T21), immobilized water (T22) and free water (T23), with the corresponding peak areas of A21, A22 and A23, respectively. During drying, the changes of A21 and A22 were not significant, while A23 and ATotal decreased significantly (p < 0.05). Linear discriminant analysis (LDA) of electronic nose data showed that samples with different drying time can be well distinguished. Hierarchical clustering analysis (HCA) confirmed that the electronic nose characteristic sensor data S4, S5, S8 and S13 corresponded with the data measured by GC-MS. The correlation analysis between LF-NMR parameters and characteristic sensors showed that A23 and ATotal were significantly correlated with the volatile components (p < 0.05). The results of the BP-ANN prediction showed that the model fitted well and had strong approximation ability (R>0.95 and error<3.65%) and stability, which indicated that the ANN model can accurately predict the flavor change during ginger drying based on LF-NMR parameters.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.