08-01-2011 06:00 PM
I am trying to replicate forecasts that were produced using the procarima function in sas. I am developing a excel workbook and using the coefficients attained from sas in excel. My problem is that I set up the equation using the dame data used by sas along with the coefficients produced by sas and I don't get the same forecast numbers. I know there is a step that sas performs that is not mentioned in the description of the arima.
When I tried to use arima to forecast a variable just using past values of the variable i had the same problem as I am having now that i am forecasting my variable of interest using past values of my variable along with other independent variables. I tried many different was of using the coefficients to produce the same forecast that i got in sas in excel. By luck i decided to subtract the constant from the past value of my variable and bam I was able to reproduce the same exact forecast that sas produced. Now that I am trying to reproduce forecast of a transfer equation I feel that once again sas performs a step that it does not mention.
Can someone help me?