Fig 1.
Support vector machine architecture.
Fig 2.
Mid-infrared spectral of 658 Cornus officinalis samples from 11 different places of origin.
Figs are generated using Matlab (Version R2021b, https://www.mathworks.com/) [Software].
Fig 3.
The average mid-infrared spectra of Cornus officinalis samples by different places of origin.
Table 1.
Principal component eigenvalues, contribution rate and cumulative contribution rate of the mid-infrared spectral data of Cornus officinalis.
Fig 4.
The number of samples in the training set, test set and all sets of Cornus officinalis from different origins.
Fig 5.
10-fold cross-validation process description and implementation.
Fig 6.
Flow chart of realization of SVM origin identification model of Cornus officinalis based on K-fold cross-validation.
Fig 7.
Confusion matrix for Cornus officinalis test samples.
Each row of the confusion matrix represents the predicted category, and each column represents the true category.
Table 2.
Precision and recall (sensitivity) of each Chinese herbal medicine origin identification model in 11 different origins, where PPV stands for precision and TPR stands for sensitivity (values are measured in %).
Table 3.
Comparison results of models for origin identification of Cornus officinalis based on mid-infrared spectroscopy.