Object Reference : Object View and Procedure Reference : Series
  
 
dsa
Seasonally adjust daily series using the DSA method.
Syntax
series_name.dsa(options) seas_name [@fa factor_name] [@trnd trend_name]
You may follow the dsa keyword with a name to save the seasonally adjusted series. Further, you may use the @fa and @trnd keywords to provide names for the saved seasonal factors and the trend series.
Options
 
forc=arg
Specify the end date of the forecast. If not specified, the last observation in the workfile is used. The forecast begins at the observation following the current workfile sample (note, if the workfile sample is equal to the workfile range, no forecasting is performed).
extendfri
For 5-day week data, interpolate to 7-day weeks by repeating the Friday value for Saturday and Sunday. Default is to perform 5-day DSA instead of converting to 7-day.
interwkend
For 5-day week data, interpolate to 7-day weeks by using linear interpolation between the Friday value and Monday values for Saturday and Sunday. Default is to perform 5-day DSA instead of converting to 7-day.
fixedarima
Use a fixed ARIMA model. Default is to use model selection to determine the ARIMA model.
nodiff
Set the level of differencing in the ARIMA model to 0. Default is 1 if using a fixed ARIMA model, or a choice between 0 and 1 if using automatic selection.
maxar=integer
If using fixed ARIMA model (see the fixedarima option), specify the AR order. If using automatic selection, specify the maximum AR order.
maxma=integer
If using fixed ARIMA model (see the fixedarima option), specify the MA order. If using automatic selection, specify the maximum MA order.
fixedtrig
Use a fixed number of trigonometric terms to model the seasonal patters in the ARIMA model. Default is to use model selection to determine the number of terms.
maxtrig=integer
If using fixed number of trigonometric terms (see the fixedtrig option), specify the number of terms. If using automatic selection, specify the maximum number of terms.
olnoao
Do not perform detection of AO outliers. Default is to detect AO outliers.
olnoio
Do not perform detection of IO outliers. Default is to detect AO outliers.
olls
Include detection of LS outliers. Default is to not detect LS outliers.
oltc
Include detection of TC outliers. Default is to not detect TC outliers.
olcvalue=arg
Specify the critical value for the outlier detection process.
oldelta=arg
Specify the delta value for the TC outlier detection process.
olinits=integer
Specify number of inner iterations in the outlier detection process.
oloutits=integer
Specify number of outer iterations in the outlier detection process.
extenddow
When forecasting day-of-week factors, repeat the last week of actual data throughout the forecast period. Default is to use exponential smoothing to forecast the factors.
prompt
Force the dialog to appear from within a program.
p
Print view.
STL options
Day-of-week
 
weeksp=integer
Specify the seasonal polynomial degree. Default is 0.
weektp=integer
Specify the trend polynomial degree. Default is 1.
weekfp=integer
Specify the filter polynomial degree. Default is 1.
weeksl=integer
Specify the length of the seasonal smoothing window (odd integers only). Default is 151.
weektl=integer
Specify the length of the trend smoothing window (odd integers only). Default is based upon the seasonal smoothing window length.
weekfl=integer
Specify the length of the filter smoothing window (odd integers only). Default is 1.
weekinits=integer
Specify number of inner iterations. Default is 1.
weekoutits=integer
Specify the number of outer iterations. Default is 15.
Day-of-month
 
monthsp=integer
Specify the seasonal polynomial degree. Default is 0.
monthtp=integer
Specify the trend polynomial degree. Default is 1.
monthfp=integer
Specify the filter polynomial degree. Default is 1.
monthsl=integer
Specify the length of the seasonal smoothing window (odd integers only). Default is 51.
monthtl=integer
Specify the length of the trend smoothing window (odd integers only). Default is based upon the seasonal smoothing window length.
monthfl=integer
Specify the length of the filter smoothing window (odd integers only). Default is 1.
monthinits=integer
Specify number of inner iterations. Default is 1.
monthoutits=integer
Specify the number of outer iterations. Default is 15.
Day-of-year
 
yearsp=integer
Specify the seasonal polynomial degree. Default is 0.
yeartp=integer
Specify the trend polynomial degree. Default is 1.
yearfp=integer
Specify the filter polynomial degree. Default is 1.
yearsl=integer
Specify the length of the seasonal smoothing window (odd integers only). Default is 13.
yeartl=integer
Specify the length of the trend smoothing window (odd integers only). Default is based upon the seasonal smoothing window length.
yearfl=integer
Specify the length of the filter smoothing window (odd integers only). Default is 1.
yearinits=integer
Specify number of inner iterations. Default is 1.
yearoutits=integer
Specify the number of outer iterations. Default is 15.
 
Example
elecdmd.dsa(forc="2015/6/30") elecdmd_adjusted
Performs daily seasonal adjustment on the ELECDMD series, specifying that the forecast end point should be 30 June 2015, and that the final adjusted series should be named ELECDMD_ADJUSTED.
elecdmd.dsa(fixedtrig, nodom, nodoy) elecdmd_adjusted @fa elecdmd_factors
Performs daily seasonal adjustment on ELECDMD, using a fixed number of trigonometric terms in the ARIMA step, and without using day-of-month or day-of-year STL. As well as saving the final adjusted series as ELECDMD_ADJUSTED, the final seasonal factor are also saved under ELECDMD_FACTORS.
Cross-references
See also Series::x12, Series::x13, and “Seasonal Adjustment”.