The TIMEDATA Procedure
- Overview
- Getting Started
-
Syntax
-
DetailsAccumulationMissing Value InterpretationTime Series TransformationTime Series DifferencingSummary StatisticsProgramming StatementsPredefined SymbolsData Set OutputOUT= Data SetOUTARRAY= Data SetOUTPROCINFO= Data SetOUTSCALAR= Data SetOUTSUM= Data Set_STATUS_ Variable ValuesPrinted OutputODS Table NamesODS Graphics Names
-
Examples
- References
Four transformations are available for strictly positive series only. Let 
be the original time series, and let 
be the transformed series. The transformations are defined as follows:
- Log
-
is the logarithmic transformation.
![\[ w_{t} = \mr {ln}(y_{t}) \]](images/etsug_timedata0004.png)
- Logistic
-
is the logistic transformation.
![\[ w_{t} = \mr {ln}(c y_{t} / (1-c y_{t})) \]](images/etsug_timedata0005.png)
where the scaling factor
is
![\[ c = (1-10^{-6}) 10 ^{- \mr {ceil}( \mr {log}_{10}({max}( y_{t}) ))} \]](images/etsug_timedata0007.png)
and
is the smallest integer greater than or equal to x.
- Square root
-
is the square root transformation.
![\[ w_{t} = \sqrt {y_{t}} \]](images/etsug_timedata0009.png)
- Box Cox
-
is the Box-Cox transformation.
![\[ w_{t} = \begin{cases} \frac{y_{t}^{{\lambda }} - 1}{\lambda }, & {\lambda } {\ne } 0 \\ \mr {ln}(y_{t}), & {\lambda } = 0 \end{cases} \]](images/etsug_timedata0010.png)
More complex time series transformations can be performed by using the SAS/ETS EXPAND procedure.