you can MAKE this happen (which you don't) by data or coding errors.
the data problems include stuff like collinearity (put the same
variable in twice and you'll find it). but in all likelihood you have
a coding error. it won't make you feel any better, but its a common
(indicator of a) problem!
Gary
---
http://gking.harvard.edu
On Sun, Feb 28, 2010 at 4:14 PM, Abigail Coplin
<abigail.coplin at gmail.com> wrote:
Hey guys,
When I try to take the inverse of the Hessian using the
command?optim.vcv<-solve(-1*my.optim$hessian), I keep getting the error msg
"Error in solve.default(-1 * my.optim$hessian) :??Lapack routine dgesv:
system is exactly singular"...any ideas how to fix this?
~Abby
_______________________________________________
gov2001-l mailing list
gov2001-l at
lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l