Regression on Airfoil Self-Noise (ASN) (k-fold cross-val)
0.0011TMSE (20% trim, 100x)rRNet
Evaluation results
| Method | Links | |
|---|---|---|
| rRNetarchitecture=3-layer MLP, hidden nodes=14, beta=0.3, k-fold=102026.02 | 0.0011 | |
| Tukey's Marchitecture=3-layer MLP, hidden nodes=14, k-fold=102026.02 | 0.0011 | |
| rRNetarchitecture=3-layer MLP, hidden nodes=14, beta=0.5, k-fold=102026.02 | 0.0012 | |
| rRNetarchitecture=3-layer MLP, hidden nodes=14, beta=0.1, k-fold=102026.02 | 0.0012 | |
| LTAarchitecture=3-layer MLP, hidden nodes=14, k-fold=102026.02 | 0.0012 | |
| rRNetarchitecture=3-layer MLP, hidden nodes=14, beta=0.7, k-fold=102026.02 | 0.0013 | |
| Huber's Marchitecture=3-layer MLP, hidden nodes=14, k-fold=102026.02 | 0.0013 | |
| MAEarchitecture=3-layer MLP, hidden nodes=14, k-fold=102026.02 | 0.0013 | |
| rRNetarchitecture=3-layer MLP, hidden nodes=14, beta=1, k-fold=102026.02 | 0.0014 | |
| LSEarchitecture=3-layer MLP, hidden nodes=14, k-fold=102026.02 | 0.0015 | |
| LTSarchitecture=3-layer MLP, hidden nodes=14, k-fold=102026.02 | 0.0015 | |
| LMLSarchitecture=3-layer MLP, hidden nodes=14, k-fold=102026.02 | 0.0015 |