Specify Hyperparameter-Tuning Values at the Plan Level

You can specify some hyperparameter-tuning values at the level of your demand or demand and supply plan and override the values for the forecasting parameters in your user-defined forecasting profile of the Bayesian machine-learning type.

Before you start

By entering values for your plan, you can exercise granular control over some settings for hyperparameter tuning. The changes you make for your plan won't apply to the forecasting profile or other plans that use the forecasting profile.

Thus, you can have multiple plans that refer to the same forecasting profile for hyperparameter tuning but use different values for some forecasting parameters for hyperparameter tuning.

Here's what to do

1. On the page for your plan, on the More Actions menu, select View More.
   The Actions drawer will open.
2. In the Plan Actions section, select Edit Plan Options.
   The guided process for editing the plan will open on a new tab of your web browser.
3. In the Demand step, on the Forecast Settings tab, in the Action column, select the icon for editing the forecasting profile for hyperparameter tuning.
   The drawer for editing the forecasting profile will open.
4. In the Hyperparameter tuning settings section, do the following:
   Note: By default, the displayed value and selections reflect the forecasting-parameter settings for the forecasting profile. The fields in this section won't be available for a predefined forecasting profile or a user-defined forecasting profile for which the HypertuneSamplePercent forecasting parameter hasn't been added with a value greater than 0 (zero).

   • Select the Enable hyperparameter tuning checkbox to turn on hyperparameter tuning for your plan.

     If this checkbox isn't selected, the other fields for hyperparameter tuning will be grayed out in the drawer.

   • In the Percentage of Nodes to Tune field, enter the percentage of forecasting nodes (combinations) in your plan for which hyperparameter tuning must be done in each plan run.

     The value displayed in this field reflects the value of the HypertuneSamplePercent forecasting parameter in the forecasting profile. The value is 0 (zero) if the Enable hyperparameter tuning checkbox is deselected.

   • In Output Options, select the Tune and use or Tune and compare option.

     When you select Tune and use, the forecast according to the tuned settings for a successfully tuned forecasting node is stored in the output measure for the forecasting profile. The tuned settings are stored in the Bookings Forecast: Tuned Settings, Shipments Forecast: Tuned Settings, or <user-defined output measure>: Tuned Settings measure.

     When you select Tune and compare, the forecast according to the tuned settings for a successfully tuned forecasting node is stored in the Bookings Forecast: Tuned, Shipments Forecast: Tuned, or <user-defined output measure>: Tuned measure, and the forecast according to the forecasting parameters for the forecasting profile is stored in the output measure. The tuned settings are stored in the Bookings Forecast: Tuned Settings, Shipments Forecast: Tuned Settings, or <user-defined output measure>: Tuned Settings measure.

     The selected option reflects the value of the HypertuneOutputMode forecasting parameter in the forecasting profile.

   • Select the Enable Planning Advisor notifications checkbox to display notifications through the Planning Advisor after plan runs when hyperparameter tuning is successful for forecasting nodes.

     The checkbox reflects the value of the HypertuneAdvisor forecasting parameter in the forecasting profile.

5. Select Update.
   The drawer will close, and you'll return to the Forecast Settings tab in the Demand step of the guided process.
6. Select Submit.
   A message confirming your updates is displayed.

What to do next

You need to run the plan to apply the changed settings for hyperparameter tuning to the forecasting nodes.

In the Run plan drawer, in the Data Refresh Options section, select the Refresh with current data option.

In the Forecasting Profiles section, select the forecasting profile for hyperparameter tuning.