Forecast Profile
Definition
In Responsive Replenishment, a forecast profile represents the collection of data that is needed when you enter forecast data in the material master record.
Use
To gain access to this object, select
Demand
StatisticalForecasting
Forecast Parameters
Forecast Profiletab
.
Structure
In any given forecast profile, the following nine forecast parameters can be defined.
Number of history periods
Specify the number of history periods to be used for forecasting.
Number of forecast periods
Specify the number of periods to be forecast.
Forecast model
Specify the forecast model – for example,
Constant
,Croston
,Season
orTrend
– to be used.Number of periods per season
Specify the number of periods per season to be used.
Alpha factor
In this field, specify a smoothing factor for the basic value.
Beta factor
In this field, specify a smoothing factor for the trend value.
Gamma factor
In this field, specify a smoothing factor for the seasonal index.
Sigma factor
In this field, specify the sigma factor to be used in calculating the tolerance lane for the historical time series, to correct outliers.
Error for model selection
In this field, specify the forecast accuracy measurement – for example,
MAD
,Error total
orMAPE
– to be used for Automatic Model Selection.
In addition to these forecast parameters, you may assign three selection options and two profiles to the forecast profile. The following three selection options can be (de)activated, if desired, by (de)selecting the corresponding checkbox.
Outlier correction
If you wish to perform outlier corrections in this forecast profile, then select this checkbox.
Without leading zero history
If you wish to prevent the leading zero history from being used during forecasting, then select this checkbox.
Use corrected history (No FDC)
If you wish to use the corrected history, rather than the original history, for forecasting, and thereby omit lifecycle planning, then select this checkbox.
Also, if desired, you may specify one or both of the following profiles within the forecast profile.
Weighting factor profile
If desired, specify a weighting factor profile, by which the values in your forecasts will be multiplied in order to give them a greater or lesser influence on the overall result.
Trend dampening profile
If desired, specify a trend dampening profile that specifies, per period the percentage by which you want the trend to be dampened.
Finally, by selecting the appropriate checkbox(es), you may define the forecast profile to calculate none, one, some, or all of the following forecast accuracy measurements:
Example
The following example provides an overview of one set of possible entries that could be defined for a forecast profile.
Screen Element |
Entry |
|---|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|