Features: 2025–2026 → Forecast: 2026–2027
24-configuration model: parameter reference
Example settings and scores for comparing training and validation error. The scores and interpretations below are the reference values provided for this table.
| max_depth | min_samples_split | min_samples_leaf | max_features | n_estimators | Train MAE | Validation MAE | Relative gap | Interpretation |
|---|---|---|---|---|---|---|---|---|
| 5 | 2 | 1 | sqrt | 200 | 0.90 | 1.05 | 14.3% | Possible underfitting |
| 10 | 5 | 2 | log2 | 300 | 0.45 | 0.72 | 37.5% | Best validation result in this example |
| 15 | 10 | 5 | 0.5 | 500 | 0.22 | 0.77 | Possible overfitting | |
| 20 | 20 | 10 | 1.0 | 300 | 0.10 | 0.85 | Strong overfitting warning | |
| None | 2 | 1 | 1.0 | 500 | 0.02 | 1.10 | Severe overfitting warning |
Only models with a generated forecast file are available to select.
Enter at least two characters. Choose a match to view its forecast.
Regression estimates
Member growth
—
DCP goals met
—
Classification estimate
Distinguished level
—
The classification label comes from its own model and is not derived from the regression estimates.