28 Feb
2010
28 Feb
'10
5:59 p.m.
Hi Meryl,
You might want to look through the lecture notes for how to approach a
question where you introduce covariates just for the mu term and not
for the sigma^2 term. My section notes from Thursday also work
through a similar example.
Maya
On Sun, Feb 28, 2010 at 3:45 PM, Meryl Federman
<federman at fas.harvard.edu> wrote:
Should our function in 1B take a value for sigma as
well, since we
are trying to optimize our log likelihood function of beta *and*
sigma? If no, how does one get the value of sigma that you need to
calculate the log likelihood?
Just concerned because it mentions our log likelihood taking a beta,
but not a sigma. (Also, if it is yet another parameter to throw into
the function, how does one optim over two variables?)
~Meryl
_______________________________________________
gov2001-l mailing list
gov2001-l at lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l