Time Series Analysis Package
Details
You can use the DECOMP object to perform seasonal decomposition (also called seasonal adjustment) on the working series. The DECOMP object uses classical decomposition. More complex seasonal decomposition or adjustment analysis can be performed by using the X11 or the X12 procedure of SAS/ETS.
The SetMode method determines the mode of the seasonal adjustment decomposition to be performed. There are four modes: multiplicative (MULT), additive (ADD), pseudo-additive (PSEUDOADD), and log-additive (LOGADD) decomposition (in which the components are exponentiated to the original metric). The default value for the SetMode method is MULTORADD, which uses multiplicative mode for series that are strictly positive, pseudo-additive mode for series that are nonnegative, and additive mode for series that are not nonnegative.
The SetLambda method specifies the Hodrick-Prescott filter parameter (Hodrick and Prescott 1980) for trend-cycle decomposition. The Hodrick-Prescott filter is used to decompose the trend-cycle component into the trend component and the cycle component in an additive fashion. The SetLambda method takes a single argument Value, which specifies the length of the seasonal cycle. A smaller Value assigns less significance to the cycle; that is, a value of 0 implies no cycle component. The default value is 1600.
Table 12 shows the notation and keywords that are associated with seasonal decomposition (adjustment) analysis.
Table 12: Seasonal Adjustment Formulas
When is odd, the trend-cycle component is computed from the
-period-centered moving average as follows:
When is even, the trend-cycle component is computed from the
-period-centered moving average as follows:
The seasonal component is obtained by averaging the seasonal-irregular component for each season,
where and
. The seasonal components are normalized to sum to 1 (multiplicative) or 0 (additive).