BJAUTOFIT issues

[ksternbe]

I am trying to run BJAUTOFIT procedures to help select a model for my series.

When I run the code below I don't get an error message but RATS does not provide any output. It simply begins to RUN for 15+ minutes and then crashes.

CALENDAR(M) 1984:1
DATA(FORMAT=FRED) * 2017:12 INDPRO
ALLOCATE 2017:12

SET LOGPRO = log(INDPRO)

*Run a BJAUTOFIT

SOURCE "Macintosh HD/Applications/RATS 9.1 Classroom/Procedures/BJAUTOFIT.src"

@BJAUTOFIT(PMAX=10,QMAX=10,CRIT=AIC) LOGPRO * 2015:12
@BJAUTOFIT(PMAX=10,QMAX=10,CRIT=SBC) LOGPRO * 2015:12

Any suggestions to deal with this issue?

Thank you.

[TomDoan]

Are you sure it's with the BJAUTOFIT? The DATA instruction is pulling data off the Internet, and is a more likely place for the program to get stuck.

However, you didn't difference an obviously non-stationary series, so a lot of the models won't fit well at all (the pure MA's for instance can't explain the rather steady trend).

[ksternbe]

I know that it is not the DATA portion of the code. In my series window I can see that my data comes in successfully. It could be an incompatible series, but I would expect an error message instead of just a 20 minute run time and then a software crash?

Thank you for your suggestions.

[ksternbe]

You were right, it worked after I took the logged first difference of the series.

Thanks again!