sm.tsa.ARMA making ma invertibility #1059

Closed
jsudheer opened this Issue Aug 21, 2013 · 1 comment

Projects

None yet

2 participants

@jsudheer

Hi,
I am using statsmodels .6 version and python 2.7.3
I was trying to fit an AR2MA2 model for my time series data and I get below error while fitting ar2ma2 model while there is no issue if ar2ma1 or ar3ma3 etc.
when I try ar2ma2 and ar3ma2 I get below error

The computed initial MA coefficients are not invertible
You should induce invertibility, choose a different model order, or you can
pass your own start_params.

Can some one advice me what exactly the issue is and how to make the MA coefficient invertible?.

I tried to do the same with R I get coefficients with out issue. So I assume R has an automatic way of making the MA coefficients invertible.
I need to stick to python for other reason as I use daily time series.
I have posted the files in the google groups.
with best regards,
Suheer

@jseabold
Member
jseabold commented Jan 8, 2014

Hi, I'm just now seeing this. There are similar discussions on the mailing list about this issue. You can read more there.

@jseabold jseabold closed this Jan 8, 2014
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment