Spreading item and location results with forecast engine Spread Measure defined and enabled
The spreading method is based on theSpreading Measure for period
constant results forecast engine's parameters and Spreading Measure for
time-phased results parameters, provided the parameters are defined and enabled
for the current cycle.
The time-pased forecast engine results of the cycle periods (history horizon or future horizon) are prorated using the Spreading Measure for time-phased results parameter. For example, Statistical Forecast DPLS_FSTAT_EXT.
The factors used in the proration is the value for each base-level child of the selected items, location and periods divided by the value for each aggregate item, location and period with a forecast value to be prorated.
For example, The forecast engine call produces forecasts for the aggregated
item node 'Car', which contains five base-level elements containing a value for the specified
spreading measure (1000, 1001, 1002, 1003 and 1004). The forecast is generated at a period
level of Months, wherein the calendar level to store the scenario value is Weeks. The ratio of
item 1000 and location CONTINENTAL, for a specific week FY16 W18 = (Value of DPLS_FSTAT_EXT for (1000, CONTINENTAL, FY16 W18)) / (Value of
DPLS_FSTAT_EXT for Car, CONTINENTAL, FY16 M05)).
Note: The FY16 M05 is the month containing
FY16 W18.
Note:
- The value of DPLS_FSTAT_EXT aggregate node must be derived from the measure values of the base-level items.
- The spreading values from the base-level elements are used to store scenario values irrespective of the number of levels that exist between the selected Forecast Engine's level and the base-level.
When prorating, these time-phased measures are interpolated using the Spreading Measure for time phased result parameter:
- Forecast
- Model fitting History
- Online Model Fit
- Retrospective Model Fit
- Seasonal Indices
- Outliers
- Step Change Exceptions
- Tracking Signal Exceptions
Note: For forecast engines, the exception
measures are time-dependent. Each period with an exception is set to 1. If applicable, the
values are prorated to the base level to store scenario values. You can view the data at the
forecasted (group or base) item, location and period level.
The time-independent Forecast Engine result values at PCONST are prorated using the Spreading Measure for
period constant results parameter. For example, Valid Combinations TUPLE_EXISTS parameter. The spreading factor is based on the
PCONST value for each base-level child of the selected item and location combination divided
by the PCONST value for each item and location level with a forecast value to be
prorated.
Note: For each item
selection, in the forecast engine call, spreading values from base-level children are used
to define the scenario values irrespective of the number of levels between the selected item
and the Cycle.item level.
For both forecast engines, the exception measures are time dependent. Each
period with an exception is set to 1. The values are stored at PCONST and are prorated using the Spreading Measure for time-phased
results parameter. You can view the data at the forecasted item and location
level.
Note: The same process can be
used for spreading results to base location when the specified location level does not match
the lowest level of the location hierarchy for the Cycle.Module.