Chapter Contents
Chapter Contents
The ARIMA Procedure

Functional Summary

The statements and options controlling the ARIMA procedure are summarized in the following table.

Description Statement Option
Data Set Options   
specify the input data setPROC ARIMADATA=
specify the output data setPROC ARIMAOUT=
include only forecasts in the output data setFORECASTNOOUTALL
write autocovariances to output data setIDENTIFYOUTCOV=
write parameter estimates to an output data setESTIMATEOUTEST=
write correlation of parameter estimatesESTIMATEOUTCORR
write covariance of parameter estimatesESTIMATEOUTCOV
write estimated model to an output data setESTIMATEOUTMODEL=
write statistics of fit to an output data setESTIMATEOUTSTAT=
Options for Identifying the Series   
difference time series and plot autocorrelationsIDENTIFY 
specify response series and differencingIDENTIFYVAR=
specify and cross correlate input seriesIDENTIFYCROSSCORR=
center data by subtracting the meanIDENTIFYCENTER
exclude missing valuesIDENTIFYNOMISS
delete previous models and start freshIDENTIFYCLEAR
specify the significance level for testsIDENTIFYALPHA=
perform tentative ARMA order identification using the ESACF MethodIDENTIFYESACF
perform tentative ARMA order identification using the MINIC MethodIDENTIFYMINIC
perform tentative ARMA order identification using the SCAN MethodIDENTIFYSCAN
specify the range of autoregressive model orders for estimating the error series for the MINIC MethodIDENTIFYPERROR=
determines the AR dimension of the SCAN, ESACF, and MINIC tablesIDENTIFYP=
determines the MA dimension of the SCAN, ESACF, and MINIC tablesIDENTIFYQ=
perform stationarity testsIDENTIFYSTATIONARITY=
Options for Defining and Estimating the Model 
specify and estimate ARIMA modelsESTIMATE 
specify autoregressive part of modelESTIMATEP=
specify moving average part of modelESTIMATEQ=
specify input variables and transfer functionsESTIMATEINPUT=
drop mean term from the modelESTIMATENOINT
specify the estimation methodESTIMATEMETHOD=
use alternative form for transfer functionsESTIMATEALTPARM
suppress degrees-of-freedom correction in variance estimatesESTIMATENODF
Printing Control Options   
limit number of lags shown in correlation plotsIDENTIFYNLAG=
suppress printed output for identificationIDENTIFYNOPRINT
plot autocorrelation functions of the residualsESTIMATEPLOT
print log likelihood around the estimatesESTIMATEGRID
control spacing for GRID optionESTIMATEGRIDVAL=
print details of the iterative estimation processESTIMATEPRINTALL
suppress printed output for estimationESTIMATENOPRINT
suppress printing of the forecast valuesFORECASTNOPRINT
print the one-step forecasts and residualsFORECASTPRINTALL
Options to Specify Parameter Values   
specify autoregressive starting valuesESTIMATEAR=
specify moving average starting valuesESTIMATEMA=
specify a starting value for the mean parameterESTIMATEMU=
specify starting values for transfer functionsESTIMATEINITVAL=
Options to Control the Iterative Estimation Process 
specify convergence criterionESTIMATECONVERGE=
specify the maximum number of iterationsESTIMATEMAXITER=
specify criterion for checking for singularityESTIMATESINGULAR=
suppress the iterative estimation processESTIMATENOEST
omit initial observations from objectiveESTIMATEBACKLIM=
specify perturbation for numerical derivativesESTIMATEDELTA=
omit stationarity and invertibility checksESTIMATENOSTABLE
use preliminary estimates as starting values for ML and ULSESTIMATENOLS
Options for Forecasting   
forecast the response seriesFORECAST 
specify how many periods to forecastFORECASTLEAD=
specify the ID variableFORECASTID=
specify the periodicity of the seriesFORECASTINTERVAL=
specify size of forecast confidence limitsFORECASTALPHA=
start forecasting before end of the input dataFORECASTBACK=
specify the variance term used to compute forecast standard errors and confidence limitsFORECASTSIGSQ=
control the alignment of SAS Date valuesFORECASTALIGN=
BY Groups   
specify BY group processingBY 

Chapter Contents
Chapter Contents

Copyright © 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.