On Sun, 10 Feb 2019, Fred Engst wrote:
here is the top part of my crash report [...]
Thanks. If we're ever to get to the bottom of what happened here, a
little more information could be helpful, namely
* What was the time-series length of your panel?
* How many regressors did you have, besides the 5000
dummies?
* Roughly how long did the estimation run before the crash?
* Was there any evidence that estimation was finished at the
time, or was it still struggling away?
Also, if you'd like to send me the full crash report offlist, maybe
I'll find something useful in it.
Allin