Does anyone know of a package (or workaround) for fitting a dirichlet
distribution by maximum likelihood?
(I'm looking for something like this: http://repec.org/bocode/d/dirifit.html,
that allows for both dependent variables summing to 1 predictive variables
of any sort.)
Don
--
Donald
On Sep 5, 2010, at 10:53 AM, Donald Braman wrote:
Does anyone know of a package (or workaround) for fitting a dirichlet
distribution by maximum likelihood?
Searching:
RSiteSearch(fitting dirichlet distribution maximum likelihood)
(Rapidly) produces this candidate:
install.packages('sos') # if you haven't already
library(sos)
dch - ???dirichlet
# found 288 matches = help pages on contributed packages
summary(dch)
# in 65 packages
dch
# lists the 288 matches in a table in a browser with links the help pages
# sorted to put the package with the most matches
3 matches
Mail list logo