FIT THE CUBIC SPLINE¶
Fits a penalised cubic spline to one or more input curves, controlling smoothness via knot count, region count, spline order, and an optional regularisation lambda. Use this worker to smooth or downsample noisy time-history or XY curves while retaining the underlying trend. K-fold cross-validation options are available to assess fit quality.
When to use¶
Classification: process.
Tagged: cross_validation, cubic_spline, curve_smoothing, downsampling, k-fold, knots, lambda, penalised_spline.
Inputs¶
| Label | ID | Type | Default | Required | Description |
|---|---|---|---|---|---|
| Curves In | curves_in | vector | — | One or more XY curves (time-history or any paired data) to be fitted with a cubic spline; accepts multiple curve objects simultaneously. | |
| Order | order | scalar | 1 | Polynomial order of the spline pieces (integer, default 1); increase to allow higher-degree local polynomials — typical values are 1–4. | |
| Number Of Knots | numberof_knots | scalar | 1 | Number of internal knot points that partition the curve domain (integer, default 1); more knots allow a closer fit but risk overfitting. | |
| Number Of Regions | numberof_regions | scalar | 1 | Number of equal-width regions used to distribute knots along the x-axis (integer, default 1); leave at 1 for uniform knot spacing. | |
| Validation | validation | scalar | none | Cross-validation scheme applied to assess spline fit quality; choose ‘none’ (default) to skip validation, or ‘k-fold-3/5/10’ for k-fold CV with the specified number of folds. | |
| Lambda | lambda | scalar | 1 | Regularisation penalty factor controlling the smoothing contribution of each knot (float, default 1.0); larger values produce smoother splines by suppressing knot influence. |
Outputs¶
| Label | ID | Type | Description |
|---|---|---|---|
| Cubic Spline Curves | curves_cs_output_1 | vector | Fitted cubic spline curves corresponding to each input curve, resampled onto the same x-axis domain with smoothed y-values. |
Disciplines¶
- ai_ml.supervised.regression
- data.curve.transform
- data.signal_processing
Runnable example¶
A runnable example is registered for this worker. Open the example workflow on the d3VIEW canvas: /api/workflow/example?id=curves_cubic_spline
Auto-generated from transformation schema. Worker id: curves_cubic_spline. Schema hash: 2ec1cd505b2d. Hand-curated docs in workerexamples/ override this page when present.