When I run an ARMA AR(1) model the output lists all of my independent variables and adds one called SIGMASQ. What is this calculating and can it be derived from the data on the printout?
Re: sigmasq
Posted: Tue Oct 13, 2015 9:27 am
by startz
This is giving the variance of the innovations. One might think it would be very close to the square of the standard error of the regression, but that doesn't seem to be the case.