MCMCjump

Prepared by Ziheng Yang, 19 April 2013

Not much is written here, but I copied the README file to hold the space here.

README.txt

Ziheng Yang, 19 April 2013

This archive includes two ANSI C programs. MCMCjump implements the discretization method and MCMC algorithms for evaluating the efficiency of various proposal kernels applied to different target distributions. The second program, MCMCtree, is from the PAML package and is for Bayesian estimation of species divergence times. The data files (mcmctree.ctl, primate6s.1st2nd.phys, primate6s.tre) are for duplicating the analyses that we did in the paper.

Windows executables are included. To run the program, start a command-line box (type cmd at the Start box), and type

MCMCjump

MCMCtree

For mac OSX or linux, you have to compile the programs. Try one of the following.

cc -o MCMCjump -m64 -O3 -funroll-loops -fomit-frame-pointer -finline-functions MCMCjump.c tools.c -lm

cc -o mcmctree -m64 -funsigned-char -O3 -funroll-loops -fomit-frame-pointer -finline-functions mcmctree.c tools.c -lm

MCMCjump

MCMCtree

References

Yang, Z and Rodríguez, CE (2013) Searching for efficient Markov chain Monte Carlo proposal kernels. PNAS in press.

Last modified, 19 April 2013