Dear All
This problem has been bugging me for quite some time. I hope someone knows a solution.
When I run an ML I sometimes use the Trace option in order to see the parameter estimates for each iteration. The reason is that the machine I use sometimes closes down. And when I have the parameters I can then easily start up the estimation again with the last parameter estimates.
My problem is that STATA seem to report at most 7 decimals in its output. This, I think, is a problem since when I use the output as starting point in a ML then the first iteration reports a lesser lnl value than the last iteration. In other words, it seems that I have lost some precision.
So my questions are:
1) do you know whether my presumption is correct?
2) and if yes, do you know how to increase the precision of STATA (trace) output?
Thank you very much in advance
Lars Geerdsen
---------------------------------------------------------------
Lars Pico Geerdsen
Herluf Trolles Gade 11
DK-1052 K�benhavn K
Denmark
Reception (+45) 33 48 08 00
Direct (+45) 33 48 09 27
Fax (+45) 33 48 08 33
Cand. polit, ph.d.
Forsker, Socialforskningsinstituttet
Ekstern Lektor, Sociologisk Institut, K�benhavns Universitet
Ph.D. in economics
Research fellow, The Danish National Institute of Social Research
Aff. Assoc. Professor, Department of Sociology, University of Copenhagen
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
