Table 6 The best-performing time window setups for the participant-wise training models
Participant | Figure of merit | All features | Time and non-linear features | Frequency features | |||
|---|---|---|---|---|---|---|---|
Overlap | No overlap | Overlap | No overlap | Overlap | No overlap | ||
P1 | Window length | 5 min | 4 min | 4, 5 min | 5 min | 5 min | 4 min |
F1 score | 100% | 100% | 100% | 90.0% | 97.5% | 100% | |
P2 | Window length | 4 min | 5 min | 5 min | 2 min | 5 min | 3 min |
F1 score | 96.8% | 92.1% | 95.5% | 90.7% | 89.4% | 92.4% | |
P3 | Window length | 4, 5 min | 5 min | 5 min | 5 min | 4 min | 4 min |
F1 score | 100% | 100% | 99.0% | 100% | 97.1% | 94.6% | |
P4 | Window length | 4, 5 min | 3 min | 4, 5 min | 5 min | 2 min | 4 min |
F1 score | 100% | 98.6% | 100% | 100% | 96.7% | 94.7% | |
P5 | Window length | 4 min | 5 min | 4, 5 min | 5 min | 3 min | 5 min |
F1 score | 99.5% | 100% | 100% | 100% | 96.4% | 100% | |
P6 | Window length | 5 min | 96.2% | 3 min | 5 min | 5 min | 4 min |
F1 score | 100% | 100% | 98.0% | 100% | 95.6% | 94.6% | |
P7 | Window length | 4, 5 min | all | 1, 4, 5 min | 5 min | 4 min | 5 min |
F1 score | 100% | 100% | 100% | 100% | 100% | 100% | |
Average | F1 score | 99.47% | 98.67% | 98.93% | 97.24% | 96.10% | 96.61% |