|
Option
|
Description
|
|---|---|
|
Forecast Settings
|
|
|
Number of
periods to forecast
|
specifies the number
of periods into the future for which multistep forecasts are made.
The larger the horizon value, the larger the prediction error variance
at the end of the horizon. By default, the horizon is 12. Valid values
are integers greater than or equal to 0 and less than 32,768.
|
|
Forecast
confidence level
|
specifies the confidence
level for the series. By default, this confidence level is 95% .
|
|
Number of
periods to hold back
|
specifies a subset of
actual time series values to hold back, starting from the end of the
last nonmissing observation. Valid values are integers greater than
or equal to 0 and less than 32,768.
|
|
Outlier Detection
Note: This option is not available
if you selected Exponential smoothing as
the forecasting model type.
|
|
|
Perform
outlier detection
|
specifies that any outliers
that are automatically detected during the creation of the model are
inputs in the model.
|