AUTOMATICALLY SELECT THE OPTIMAL POLYNOMIAL ORDER USING R2¶
Automatically selects the lowest-order polynomial that achieves a target R² goodness-of-fit for a given curve, using binary search up to a configurable maximum order. Use this worker to determine the most parsimonious polynomial representation of a raw curve before downstream fitting or surrogate-model steps.
When to use¶
Classification: process.
Tagged: binary_search, curve_fitting, goodness_of_fit, model_selection, optimal_order, polynomial, r2, regression.
Inputs¶
| Label | ID | Type | Default | Required | Description |
|---|---|---|---|---|---|
| Base Curve | base_curve | vector | — | The reference or raw XY curve to be fitted; supplied as a vector of coordinate pairs — required in practice even though marked optional, as it is the curve against which all polynomial orders are evaluated. | |
| Target R2 | target_r2 | string | 0.999 | R² threshold (dimensionless, 0–1) that the fitted polynomial must meet or exceed; the worker selects the lowest order satisfying this threshold, defaulting to 0.999 when omitted. | |
| Max Order | max_order | string | 10 | Upper bound (positive integer) on the polynomial degree to consider during binary search; orders from 1 to this value are evaluated, defaulting to 10 when omitted. |
Outputs¶
| Label | ID | Type | Description |
|---|---|---|---|
| Selected Polynomial Order | selected_order | keyvalue | Key-value record containing the chosen polynomial degree — the lowest order whose R² meets the target, or the best-found order if the target was never reached. |
| R2 At Selected Order | selected_r2 | keyvalue | Key-value record reporting the R² score (dimensionless, 0–1) of the polynomial fit at the selected order. |
| Fitted Polynomial Curve | fitted_curve | vector | XY vector of the polynomial curve evaluated at the original abscissa points of the base curve, using the selected optimal order. |
| Order vs R2 Evaluation Table | evaluation_table | dataset | Tabular dataset listing each candidate polynomial order alongside its corresponding R² value, enabling inspection of the full order-vs-fit-quality trade-off. |
Disciplines¶
- ai_ml.model_selection
- ai_ml.supervised.regression
- data.curve.transform
Auto-generated from transformation schema. Worker id: curve_get_optimal_polynomial_order. Schema hash: b0e25959df34. Hand-curated docs in workerexamples/ override this page when present.