@yahoo0806 wrote:
Hi all,
I am new to proc MI procedure. I just want to ask when you see the following message, what direction are you going to consider in order to fix your code or if you can let me know, based on your experience, what can be wrong if you see the below message:
WARNING: The EM algorithm (MLE) fails to converge after 200 iterations. You can increase the number of iterations
(MAXITER= option) or increase the value of the convergence criterion (CONVERGE= option).
WARNING: The EM algorithm (posterior mode) fails to converge after 200 iterations. You can increase the number of
iterations (MAXITER= option) or increase the value of the convergence criterion (CONVERGE= option).
ERROR: Not enough observations to fit regression models in monotone-data posterior step.
Many thanks!
Tina
I don't use this proc often. The first bit in the WARNING shows things that you could edit to possibly get things to work. As it says either increase the value of the MAXITER option or the Converge option, which would be in the EM statement. If you didn't provide an EM statement then the proc used MAXITER=200 and Converge= 1E-4. So trying MAXITER=400 might be appropriate, or Converge= 1E-3 (0.001) might be reasonable.
However the Error raises the question of how many observations are in the data set and how many variables are you trying to impute?
You may have too many missing values for the number of observations.
It is a good idea to copy the entire procedure or data step along with any notes, messages and errors from the log and paste the whole thing into a code box opened on the forum with the {I} icon. That way we see all of the options you actually used. It may be that you provided a combination of options that are causing problems with a smallish sample.
