Table 4.
Classification confusion matrix (Participant A).
| Model | Predicted class | Target class | ACCclass | FPR | FNR | ||||
|---|---|---|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | |||||
| LSSVM1 | 1 | 168 | 1 | 0 | 0 | 0 | 0.9438 | – | – |
| 2 | 10 | 162 | 86 | 51 | 4 | 0.9153 | 0.0562 | 0.0028 | |
| 3 | 0 | 14 | 8 | 30 | 6 | 0.0851 | 0.0394 | 0.7622 | |
| 4 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | |
| 5 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | |
| LSSVM2 | 1 | 177 | 1 | 0 | 0 | 0 | 0.9944 | – | – |
| 2 | 1 | 174 | 17 | 0 | 0 | 0.9831 | 0.0056 | 0.0028 | |
| 3 | 0 | 2 | 75 | 50 | 0 | 0.7979 | 0.0056 | 0.0919 | |
| 4 | 0 | 0 | 2 | 30 | 7 | 0.3704 | 0.0045 | 0.5495 | |
| 5 | 0 | 0 | 0 | 1 | 3 | 0.3000 | 0.0019 | 0.7000 | |
The number of correctly and wrongly classified data is shown on the main and off diagonal, respectively. The last three columns present the classification rate of each class and the FPR and FNR for binary classification, respectively. The higher classification accuracy and lower FPR and FNR are marked in bold.