Table 3 Model correct classification rate comparison results from five-fold cross-validation of 5 salmon groups, and the origin authenticity identification results of 17 test samples by using created model (6 replicants of each sample)
From: Data fusion and multivariate analysis for food authenticity analysis
Model | Group | Alaska | Iceland_Farmed | Iceland_Wild | Norway | Scotland | Outlier | Correct rate (%) |
---|---|---|---|---|---|---|---|---|
KNN | Alaska | 57 | 1 | 15 | 20 | 6 | 0 | 85.47 |
Iceland_Farmed | 0 | 47 | 0 | 3 | 0 | 0 | ||
Iceland_Wild | 0 | 1 | 89 | 0 | 0 | 0 | ||
Norway | 0 | 0 | 0 | 100 | 0 | 0 | ||
Scotland | 0 | 1 | 5 | 24 | 153 | 0 | ||
Test samples | 0 | 0 | 0 | 0 | 0 | 102 | 0 | |
LDA | Alaska | 99 | 0 | 0 | 0 | 0 | 0 | 100 |
Iceland_Farmed | 0 | 50 | 0 | 0 | 0 | 0 | ||
Iceland_Wild | 0 | 0 | 90 | 0 | 0 | 0 | ||
Norway | 0 | 0 | 0 | 100 | 0 | 0 | ||
Scotland | 0 | 0 | 0 | 0 | 183 | 0 | ||
Test samples | 25 | 1 | 1 | 1 | 35 | 39 | 58.83 | |
RF | Alaska | 99 | 0 | 0 | 0 | 0 | 0 | 100 |
Iceland_Farmed | 0 | 50 | 0 | 0 | 0 | 0 | ||
Iceland_Wild | 0 | 0 | 90 | 0 | 0 | 0 | ||
Norway | 0 | 0 | 0 | 100 | 0 | 0 | ||
Scotland | 0 | 0 | 0 | 0 | 183 | 0 | ||
Test samples | 41 | 1 | 1 | 0 | 53 | 6 | 92.16 | |
SVM | Alaska | 96 | 0 | 1 | 0 | 2 | 0 | 98.64 |
Iceland_Farmed | 0 | 47 | 0 | 0 | 3 | 0 | ||
Iceland_Wild | 0 | 0 | 90 | 0 | 0 | 0 | ||
Norway | 0 | 0 | 0 | 99 | 1 | 0 | ||
Scotland | 0 | 0 | 0 | 0 | 183 | 0 | ||
Test samples | 22 | 1 | 2 | 3 | 74 | 0 | 94.12 | |
PLS-DA | Alaska | 99 | 0 | 0 | 0 | 0 | 0 | 100 |
Iceland_Farmed | 0 | 50 | 0 | 0 | 0 | 0 | ||
Iceland_Wild | 0 | 0 | 90 | 0 | 0 | 0 | ||
Norway | 0 | 0 | 0 | 100 | 0 | 0 | ||
Scotland | 0 | 0 | 0 | 0 | 183 | 0 | ||
Test samples | 42 | 0 | 0 | 0 | 60 | 0 | 100 | |
OPLS-DA | Alaska | 99 | 0 | 0 | 0 | 0 | 0 | 100 |
Iceland_Farmed | 0 | 50 | 0 | 0 | 0 | 0 | ||
Iceland_Wild | 0 | 0 | 90 | 0 | 0 | 0 | ||
Norway | 0 | 0 | 0 | 100 | 0 | 0 | ||
Scotland | 0 | 0 | 0 | 0 | 183 | 0 | ||
Test samples | 42 | 0 | 0 | 0 | 60 | 0 | 100 |