Path Loss Prediction on Outdoor environment dataset
4.83MAEKAN
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| KANExpression=Learned tree and splines as in Figure 132025.05 | 4.83 | 50.54 | 4.66 | 0.57 | — | — | |
| DSR-PQTExpression=log10(d / hED) + d / 10 + 902025.05 | 7.38 | 89.02 | 7.2 | 0.33 | — | — | |
| TabNetExpression=Feature importance: hED: 0.2566, d: 0.5952, f: 0.14812025.05 | 7.44 | 90.69 | 7.21 | 0.23 | — | — | |
| DSR-RSPGExpression=hEDd + log10(f) + 802025.05 | 7.69 | 95.6 | 7.52 | 0.28 | — | — | |
| KAN Auto-symbolicExpression=cos (8.7f − 3.5) − 2.3 log10 (7.7 − 2.2hED) + 105.82025.05 | 7.73 | 99.86 | 7.53 | 0.15 | — | — | |
| ResNet-MLPExpression=/2025.05 | 7.87 | 101.74 | 7.48 | 0.14 | — | — | |
| DSR-VPGExpression=(d + sin(f)) / 10 + 902025.05 | 8.02 | 101.66 | 7.87 | 0.23 | — | — | |
| PL EO Eq. 5*Expression=10n log10(d) + PL0 + Lh log10(hED) + Xσ2025.05 | 18.09 | 398.88 | 17.82 | 0.3 | — | — | |
| PL FS Eq. 7*Expression=20 log10(d) + 20 log10(f) + 32.442025.05 | 34.08 | 1,242.99 | 31.43 | 0.3 | — | — |