Reserved Variable Names

For each project, SAS Visual Forecasting creates several output data sets. The variable names in your input data set cannot match any of the variable names in these output data sets. The variable names in your input data set also cannot start with an underscore. If you try to assign a variable to a role and the variable name matches either of these conditions, then an error message appears.

The following table lists alphabetically the variables that are used by SAS Visual Forecasting.

Reserved Variable Names
Variable Name	Description
_VariableName	Variables in your project should not start with an underscore
AADJRSQ	Amemiya’s adjusted R-square
ACTUAL	Dependent series value
ADJRSQ	Adjusted R-square
AGGCHILDPREDICT	Aggregated prediction of child nodes
AIC	Akaike Information Criterion
AICC	Finite sample corrected Akaike Information Criterion
APC	Amemiya’s Prediction Criterion
CUBIC	Predefined variable for a cubic trend
DFE	Degrees of freedom error
END	Ending value of the time variable
ENDOBS	Number of the last observation in the data
EQUALITY	Equality constraint for the predicted reconciled value
ERROR	Prediction errors
FINALPREDICT	Predicted value for the parent node
GMAPE	Geometric mean percent error
GMAPPE	Geometric mean predictive percent error
GMAPES	Geometric mean absolute error percent of standard deviation
GMASPE	Geometric mean symmetric percent error
GMRAE	Geometric mean relative absolute error
IMASE	In-sample mean absolute scaled error
INVERSE	Predefined variable for an inverse trend
ISRECONCILED	If the node is reconciled, then this variable is 1; If the node is not reconciled, then this variable is 0.
LEAF	Keyword used in model generation
LINEAR	Predefined variable for a linear trend
LOWBFOVR	Lower confidence limits before override reconciliation
LOWER	Lower confidence limits
LOWERBD	Lower bound on the forecast
MAE	Mean absolute error
MAPE	Mean absolute percent error
MAPES	Mean absolute error percent of standard deviation
MAPPE	Symmetric mean absolute predictive percent error
MASE	Mean absolute scaled error
MAX	Maximum value
MAXAPES	Maximum absolute error percent of standard deviation
MAXERR	Maximum error
MAXPE	Maximum percent error
MAXPPE	Maximum predictive percent error
MAXRE	Maximum relative error
MAXSPE	Maximum symmetric percent error
MDAPE	Median percent error
MDAPES	Median absolute error percent of standard deviation
MDAPPE	Median predictive percent error
MDASPE	Median symmetric percent error
MDRAE	Median relative absolute error
ME	Mean error
MEAN	Mean value
MIN	Minimum value
MINAPES	Minimum absolute error percent of standard deviation
MINERR	Minimum error
MINPE	Minimum percent error
MINPPE	Minimum predictive percent error
MINRE	Minimum relative error
MINSPE	Minimum symmetric percent error
MPE	Mean percent error
MPPE	Mean predictive percent error
MRAE	Mean relative absolute error
MRE	Mean relative error
MSE	Mean square error
MSPE	Mean symmetric percent error
N	Number of nonmissing observations or number of variance products
NAME	Variable name
NMISS	Number of missing observations
NMISSA	Number of missing actuals
NMISSP	Number of missing predicted
NOBS	Number of observations
NONMISSCHLD	Number of nonmissing children in the current AGGBY group
NPARMS	Number of model parameters
PREBFOVR	Predicted values before override reconciliation
PREDICT	Predicted values
RECDIFF	Reconciliation difference
QUADRATIC QUAD	Predefined variable for a quadratic trend
RMSE	Root mean square error
RMSSE	Root mean square scaled error
RSQUARE	R-square
RWRSQ	Random walk R-square
SBC	Schwarz Bayesian information criterion
SEASONAL	Predefined variable for seasonal dummies
SMAPE	Symmetric mean absolute percent error
SSE	Sum of squares error
SST	Corrected total sum of squares
START	Beginning value of the time variable
STARTOBS	Number of the first observation
STD	Prediction standard errors
STDBFOVR	Standard deviation before override reconciliation
STDDEV	Standard deviation
SUM	Summation value
TOP	Keyword used in model generation
TSS	Total sum of squares
UMSE	Unbiased mean square error
UNLOCK	For locked overrides, the value of this variable is 0. For unlocked overrides, the value is 1.
UPPBFOVR	Upper confidence limits before override reconciliation
UPPER	Upper confidence limits
UPPERBD	Upper bound on the forecast
URMSE	Unbiased root mean square error
Y	Represents the dependent variable

Last updated: March 16, 2026