Kontakt

Ansicht
Dokumentation

Forecast Profiles ( RELNAPO_30A_SP1_FCS_FCPR )

Forecast Profiles ( RELNAPO_30A_SP1_FCS_FCPR )

CL_GUI_FRONTEND_SERVICES - Frontend Services   SUBST_MERGE_LIST - merge external lists to one complete list with #if... logic for R3up  
This documentation is copyright by SAP AG.
SAP E-Book

Short text

Forecast Profiles

Scope of Functions

Planning contains a number of changes and additions to the forecast profiles.

  • Master Forecast Profile
  • Time buckets profile

You must now enter a planning buckets profile in the master forecast profile. This planning buckets profile must contain only one periodicity (for example, months only). The sum of the periods in the planning buckets profile must not be smaller than the combined sum of the periods in the forecast and historical horizons that you also define in the master forecast profile.
  • Factory calendar

The forecast can be based on a factory calendar. This entry is optional.
In Release 2.0A you defined the factory calendar in the InfoCube, which meant that all forecasts in this InfoCube had to be based on the same factory calendar. Now, you can use different factory calendars within the same planning area.
  • Alert restriction

This field used to be called Alert profile. The field has the same function as in Release 2.0A, except that you now define the alert restriction (Goto -> Alert restriction) for a planning area, not for an InfoCube.
  • Offset

You can now define offsets for the forecast horizon and/or the historical horizon in the master forecast profile. +8 in the Offset field (forecast horizon) means that the forecast is calculated starting on the 9th period into the future.
To use this option, you must define the forecast horizon by entering a number of periods. Do not enter From and To dates.
  • Univariate Forecast Profile
  • Mark time series (Historical value markings as of Support Package 2)

If you have selected Outlier correction, you can enter the key of a set of historical value markings here. The markings are then applied to the historical input key figure as defined in the univariate forecast profile. They specify which of the historical values are to be included in the outlier correction function. The markings are saved on the database as a time series under this key. If the marked values lie outside the tolerance lane, they are corrected. If you make no entry in this field, all the historical values are included in the outlier calculation. You can also create or change a set of historical value markings on this screen by clicking on the Maintain historical value markings icon next to this field.
Alternatively, you can mark the time series values on the Forecast view of interactive demand planning. This is the recommended procedure. The following settings affect how historical data is corrected to account for promotions.
  • Promotion

In this field you enter a key figure in which a past promotion or promotions are stored. History is corrected by the amount of the promotions. This field is used in conjunction with one of the following checkboxes.
To use this option you must:
(i) Check this box.
(ii) Select Outlier correction.
(iii) Make an entry in the field Mark time series (see above).
The system marks all the periods within the horizon of the promotion in the past and performs outlier correction for these periods. Any historical values that lie outside the tolerance lane are corrected.
  • Change values

If you check this box, the system corrects the historical data by deducting from it the promotions.
  • Multiple Linear Regression Profile
  • Measured value error

This field defines the quality of the historical values of the dependent variable in the MLR model.
- NORMCONST: A normal distribution is used with a constant standard deviation. You can enter the constant in the field Sigma. If you do not make an entry in the field Sigma, the system uses a default of 1.
- NORMVARIO: A normal distribution is used with a variable standard deviation. The system calculates the deviation. You make no entry in the field Sigma. Use this option if you have outliers.
- SQRT: The standard deviation is the root of the measured value (that is, the root of the historical value).
- Blank: An entry in this field is optional. If you leave the field blank, the standard deviation is twice the root of the measured value (that is, twice the root of the historical value).
  • Sigma

If you have selected NORMCONST as the measured value error (see above), you can make an entry in this field. The forecast values are not affected by the value of sigma because the NORMCONST option assumes that sigma is constant and equal for all data points. However, the covariance does change, and therefore all measures of fit that are based on the covariance such as Durbin-h.





rdisp/max_wprun_time - Maximum work process run time   General Material Data  
This documentation is copyright by SAP AG.

Length: 6093 Date: 20240425 Time: 092229     sap01-206 ( 100 ms )