.. _auto_dataset_reconstruct_curve_from_predictions:

*DATASET RECONSTRUCT CURVE FROM PREDICTIONS*
============================================

Reconstructs full curves from a set of predicted discrete points by interpolating against an original reference dataset. Supports fit, linear, and ordered interpolation strategies, with optional X-axis synchronization. Use this worker to recover time-history or spatial curves from ML model predictions that output only sampled point values.

When to use
-----------

Classification: **process**.

Tagged: ``curve_reconstruction``, ``fit``, ``interpolation``, ``linear``, ``ml_postprocessing``, ``ordered``, ``predictions``, ``sync_x``.

Inputs
------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20 20 20

   * - Label
     - ID
     - Type
     - Default
     - Required
     - Description
   * - Original Dataset
     - dataset_1
     - dataset
     - —
     - 
     - Original reference dataset containing the source curves used as the interpolation basis; each row should correspond to a sample and include at least the curve column specified by 'raw_column_name'.
   * - Original Curve Column
     - raw_column_name
     - scalar
     - —
     - 
     - Name of the column in the original dataset that holds the reference curve objects or encoded curve values; must match a column present in dataset_1 (dynamically populated from dataset_1).
   * - Prediction Dataset
     - dataset
     - dataset
     - —
     - 
     - Prediction dataset whose rows contain the discrete predicted point values (columns matched by the 'prediction_points' prefix) that will be used to reconstruct the full curves.
   * - Prediction Points
     - prediction_points
     - scalar
     - Y_
     - 
     - Column-name prefix (default 'Y_') used to identify the predicted-point columns in the prediction dataset; all columns whose names start with this prefix are treated as ordered Y-values of the reconstructed curve.
   * - Type Of Reconstruction
     - reconstruction_type
     - select
     - fit
     - 
     - Interpolation strategy for curve reconstruction: 'fit' applies a best-fit interpolation, 'linear' uses piecewise linear interpolation, and 'ordered' maps predicted points directly in sequence; defaults to 'fit'.
   * - Sync X-values
     - sync_x_values
     - select
     - yes
     - 
     - Whether to synchronize the reconstructed curve's X-axis values with those of the original reference curves ('yes') or use the predicted dataset's own X coordinates ('no'); defaults to 'yes'.
   * - X Prediction Points
     - x_prediction_points
     - scalar
     - X_
     - 
     - Column names that contains the x-predicted points

Outputs
-------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20

   * - Label
     - ID
     - Type
     - Description
   * - dataset_reconstruct_curve_from_predictions_output_1
     - dataset_reconstruct_curve_from_predictions_output_1
     - dataset
     - Output dataset containing one reconstructed curve per input prediction row, with X-values and interpolated Y-values combined into a curve column aligned to the original reference domain.

Disciplines
-----------

- ai_ml.supervised.regression
- data.curve.pair
- data.curve.transform

Runnable example
----------------

A runnable example is registered for this worker. Open the example workflow on the d3VIEW canvas: `/api/workflow/example?id=dataset_reconstruct_curve_from_predictions <https://www.d3view.com/api/workflow/example?id=dataset_reconstruct_curve_from_predictions>`_

.. raw:: html

   <hr style="margin-top:2em">
   <p style="font-size:11px;color:#888">
   Auto-generated from <code>transformation</code> schema. Worker id: <code>dataset_reconstruct_curve_from_predictions</code>. Schema hash: <code>c563048b0117</code>. Hand-curated docs in <code>workerexamples/</code> override this page when present.
   </p>