Back to models
Transformer

TopdownReconciler

API referenceSource
Inverse transformUnequal-length series

Apply Topdown hierarchical reconciliation.

Forecast proportions keep the original series during transform, and propagate the “proportions” of each forecast with respect to its total during inverse_transform.

Topdown share, on the other hand, transforms the series to share the forecast with respect to their parent, and then uses the total forecast to multiply the shares.

For more information, see “Single level approaches” in [1].

Quickstart

python
from sktime.transformations.hierarchical.reconcile import TopdownReconciler

estimator = TopdownReconciler(method='td_fcst')

Tags

Capabilities

  • Inverse transform: Supported
  • Unequal-length series
  • Multivariate: Not supported
  • Missing values: Not supported
  • Removes missing values: Not supported
  • Equalizes series length: Not supported

Properties

Input typescitype:transform-input
Series
Output typescitype:transform-output
Series
Label typescitype:transform-labels
None
Fit is emptyfit_is_empty
No
Keeps the time indextransform-returns-same-time-index
No
Requires Xrequires_X
Yes
Requires yrequires_y
No
X and y need the same indexX-y-must-have-same-index
No

Parameters(1)

methodstr, default=”td_fcst”
The method to use for reconciliation. - td_fcst: Forecast Proportions. - td_share: Topdown Share.

Examples

>>> from sktime.transformations.hierarchical.reconcile import (
... TopdownReconciler)
>>> from sktime.utils._testing.hierarchical import _make_hierarchical
>>> from sktime.forecasting.naive import NaiveForecaster
>>> y = _make_hierarchical ()
>>> pipe = TopdownReconciler () * NaiveForecaster ()
>>> pipe = pipe. fit (y)
>>> y_pred = pipe. predict (fh = [1, 2, 3 ])

References

[1]

Hyndman, R. J., & Athanasopoulos, G. (2018). Forecasting: principles and practice. OTexts.