Autonomous rollout classification on Dataset E sub-cl. MoG
99.63Mean Training AccuracyKOMET
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| KOMETDrift type=sub-cl. MoG, Classes / nθ=3/27, Koopman basis=Fourier, PCs / nθ=7/27, Spectral radius ρ(A)=0.8747, Rollout steps=100, Zero retraining=true2026.03 | 99.63 | 96.97 | 99.65 | 97.98 | 0 | — | — | — | 0.572 | 54 | 1.22 | |
| FrozenDrift type=sub-cl. MoG, Weights=t=299 weights unchanged throughout2026.03 | — | — | — | — | — | — | 33.79 | 77 | — | — | — | |
| Retrained (upper bound)Drift type=sub-cl. MoG, Mode=Fresh retraining at every prediction step2026.03 | — | — | — | — | — | 99.64 | — | — | — | — | — |