Table 3.
Results of classification experiment 2 (CE2), including the performance of the 2D CNN trained in experiment 1 (CE1), on dataset 4’s test set with more weak positive results (N = 237).
| Model/heuristic/participants | Specificity | Sensitivity | Accuracy | Cohen’s kappa |
|---|---|---|---|---|
| 2D CNN (CE1) | 1.000 | 0.852 | 0.949 | 0.883 |
| 2D CNN (CE2, retrained) | 0.987 | 0.901 | 0.958 | 0.905 |
| 1D CNN | 0.968 | 0.844 | 0.926 | 0.733 |
| Heuristic-C | 0.976 | 0.487 | 0.815 | 0.525 |
| ViT Net | 0.968 | 0.952 | 0.962 | 0.917 |
| Study Participants | 0.974 | 0.679 | 0.873 | 0.699 |
CNN Convolutional Neural Network.