Restructure: documented layout, tuned results, validation artifacts, camera-ready paper
f868041 verified | model,rmse_default,mae_default,r2_default,rmse_tuned,mae_tuned,r2_tuned,rmse_ci95_low,rmse_ci95_high,tuned_params | |
| Persistence,0.1381,0.0879,-0.7655,0.1381,0.0879,-0.7655,0.1244,0.1499,{} | |
| MLP (PyTorch),0.115,0.0767,-0.2238,0.1049,0.069,-0.0182,0.0976,0.1115,"{""hidden"": [64, 32], ""lr"": 0.003, ""wd"": 0.0001, ""drop"": 0.2}" | |
| Ridge,0.097,0.0635,0.1289,0.0955,0.0617,0.1558,0.0873,0.1024,"{""alpha"": 100.0}" | |
| Random Forest,0.0992,0.0642,0.0879,0.095,0.0613,0.165,0.0864,0.1025,"{""n_estimators"": 500, ""max_depth"": null, ""min_samples_leaf"": 10, ""max_features"": ""sqrt""}" | |
| Gradient Boosting,0.0964,0.0638,0.1395,0.0946,0.062,0.1717,0.0867,0.1015,"{""n_estimators"": 800, ""learning_rate"": 0.01, ""max_depth"": 3, ""subsample"": 0.8}" | |