Function forecast

forecast( measure, name?, options?): CalculatedMeasure

Creates a calculated measure that generates a forecast based on a specified measure employing advanced autoML techniques.

This function offers flexible options allowing you to:

• Let the function auto-select the best statistical model or explicitly choose a preferred model
• Control the time period used for training the model
• Set additional options to improve forecast accuracy by supplying expected lower and upper limits.

In addition to the forecast, upper and lower confidence intervals are returned with the name of the new measure and a suffix of _upper and _lower respectively.

Forecast requires a Sisense instance version of L2023.6.0 or greater.

Parameters

Parameter	Type	Description
measure	Measure	Measure to apply the forecast logic to
name?	string	Name for the new measure
options?	ForecastFormulaOptions	Forecast options

Returns

CalculatedMeasure

A calculated measure instance

Example

Calculate a forecast for total cost from the Sample Ecommerce data model from the Sample Ecommerce data model.

measureFactory.forecast(
  measureFactory.sum(DM.Commerce.Cost),
  'Total Cost Forecast',
  {
    modelType: 'prophet',
    roundToInt: true,
  }
)