Table 4.
The results of the classification models based on fusion strategies.
| Fusion Strategy | Models | Cal | Val | Pre |
|---|---|---|---|---|
| measurement fusion | SVM | 99.903 ± 0.069 b | 99.147 ± 0.378 b | 99.535 ± 0.324 a |
| LR | 99.852 ± 0.029 b | 99.612 ± 0.137 a | 99.690 ± 0.402 a | |
| CNN | 99.981 ± 0.018 a | 99.845 ± 0.087 a | 99.922 ± 0.174 a | |
| PCA feature fusion | SVM | 99.897 ± 0.058 b | 99.186 ± 0.373 b | 99.535 ± 0.324 a |
| LR | 99.987 ± 0.029 a | 99.961 ± 0.087 a | 99.767 ± 0.318 a | |
| CNN | 99.936 ± 0.076 ab | 99.031 ± 0.565 b | 99.457 ± 0.162 a | |
| deep feature fusion | SVM | 100.000 ± 0.000 a | 99.922 ± 0.174 a | 99.922 ± 0.106 a |
| LR | 99.987 ± 0.018 a | 99.845 ± 0.162 a | 99.922 ± 0.106 a | |
| CNN | 99.981 ± 0.029 a | 99.884 ± 0.260 a | 99.922 ± 0.106 a |
The letters (a, b) in each column indicate the significance of the difference in the accuracy of different models at the confidence level of 5%. Within a column, data followed by different letters are significantly different.