Table S6. Results of the WF-CNN classifier for post-HI spike transient identification in experimental data (entire 6 hours – 7 layers)
Trained and validated on Sheep No. | No. of patterns in the train dataset | Tested on Sheep No. | No. of patterns in the Test-set | TP hits | TN hits | FP hits | FN hits | Sensitivity [%] | Selectivity [%] | Precision [%] | Accuracy [%] |
2,3,4,5,6,7 | 4567 | 1 | 443 | 170 | 270 | 3 | 0 | 100 | 98.3 | 100 | 99.3 |
1,3,4,5,6,7 | 4751 | 2 | 259 | 108 | 149 | 2 | 0 | 100 | 98.2 | 100 | 99.2 |
1,2,4,5,6,7 | 4731 | 3 | 279 | 82 | 179 | 1 | 17 | 82.8 | 98.8 | 91.3 | 93.5 |
1,2,3,5,6,7 | 3372 | 4 | 1638 | 821 | 783 | 3 | 31 | 96.4 | 99.6 | 96.2 | 97.9 |
1,2,3,4,6,7 | 4088 | 5 | 922 | 454 | 455 | 13 | 0 | 100 | 97.2 | 100 | 98.6 |
1,2,3,4,5,7 | 4466 | 6 | 544 | 231 | 309 | 0 | 4 | 98.3 | 100 | 98.7 | 99.3 |
1,2,3,4,5,6 | 4085 | 7 | 925 | 208 | 705 | 1 | 11 | 95.0 | 99.5 | 98.7 | 98.7 |
Overall performance of the 7 layers WF-CNN in the entire 6 hours | 98.07±1.92 |