Re: Gamma distribution & Markov chain Monte Carlo

Thank you very much for your comment. The reason I use MCMC is just
for testing and I believe you are completely right about other faster
methods.

Based on your comment, I assume that my code is right (if w/o burn-in)
but I'm still in serious doubt. I removed the burn-in steps as you say
but the problem remains the same, even when I tried to run for several
times. In addition, all variable values are smaller than zero, none is
bigger.

Can you please skim through my code and see if there are some
mistakes?

Thank you very much,
Iaoh.


The Gamma distribution is log concave if the shape parameter is
at least one; the log Gamma distribution for any value of
the parameter. If the parameter is at least 1/3, the cube
root of a Gamma random variable is "dominated" by the normal
density fitting it at the mode; this is the Wilson-Hilferty
transformation.  Acceptance rejection methods are quick.

MKMC does not give independent Gamma random variables; this
does with no "burn in".


.

Loading