mcmc

MCMC -- Markov Chain Monte Carlo Tools

Description

MCMC -- Markov Chain Monte Carlo Tools Copyright (c) 1998, Harvard University. Full copyright in the file Copyright. Some routines require version 5 of Matlab, some will work with version 4

 There are three parts to this library of routines.
 1. *[rnd,pdf,lpr].m - distribution function tools to complement MATLAB's
 2. mcmc*.m - routines to calculate and display summaries of MCMC output
 3. other - other useful routines
  
   1. Distribution Function Tools

   These function help in random number generation and various calculations involving density functions.

 randrand - randomize both random number chains off the clock
 mvnormrnd - random multivariate normal - different from MATLAB's mvnrnd
 wishrnd - random Wishart value
 wishirnd - random Wishart value - integer df only
 invwishrnd - random inverse Wishart value
 invwishirnd - random inverse Wishart value - integer df only
 invwishpdf - inverse Wishart density
 metrop - a general Metropolis-Hastings step
 betalpr - log probability ratio for beta distribution
 gamlpr - log probability ratio for gamma distribution
 invwishlpr - log probability ratio for inverse wishart distribution
 mvnormlpr - log probability ratio for multivariate normal distribution
   
   2. MCMC Summaries
     
 mcmclt - lower triangle - for symmetric matricies to use with mcmctrace
 mcmcsumm - calculate summary statistics(includes autocorrelations and Gelman-Rubin statistics)
 mcmctrace - matrix of trace plots
 mcmcacf - to plot autocorrelations
 mcmcgr - Gelman-Rubin R statistic for convergence
 mcmcdemo - short demonstration program
 
   3. Other

 ltvec - convert a lower-triangular matrix into a vector
 veclt - convert a vector into a lower-triangular matrix
 ltindex - convert row and column index into lt-index

Please see the contents.m file for a full description.

• MATLAB release: MATLAB 5.2 (R10)

[Erasmo Mejia]

 

[pulak Ghosh]

 

[Birajashis Pattnaik]

Problem with vesion

[Peter Xu]

Could you send me the examples of MCMC (Markov Chain Monte Carlo) modelling.

Thanks.

from Peter Xu

[mohammad atef]

plz send the file to me

[Felipe Reyes]

 

[Saverio Massi]

Very good random generator for the Wishart: works much better than the one included in matlab (which is sometimes returning "strange samples")

[Mario Albeerto]

inverse log function

[chen bw]

I don't know how to excute ARS (adaptive rejection sampling) with matlab if the density is not standard.

[zhao xiaolin]

 

[la chen]

 

[yao hs]

 

[Atique Malik]

 

[Estevan Charreteiro]

The veclt command only works for column vectors. But one can but in a transformation step to suit needs. Still a good program. thanks for posting.

[Zhang J.L.]

Your programs are helpful for me, save my time. I will study it,thank you!

[kadek nanax]

 

[ipek cameller]

 

[Zuofeng]

Hi David,

Thanks very much for help!

I am using your matlab function "mcmcgr" and have found one question.

I set up 5 chains each with length say 10000, and each chain only contains
zero components, so each chains is like "0,0,0,..." with 10000 zeros in it.

Obvious the convergence holds, but after I used mcmcgr to find its R
statistics, a "NaN" was reported. I don't know why this happens.

I am very grateful if you can help me with it!
Thanks a lot!

[David]

There is a bug in this code:

In the definition of of the inverse wishart pdf the multivariate gamma distributrion is calculated as gammaln((d-k-klst)/2) which does not fit the definition of the multivariate gamma distribution: it should be gammaln(d - (k-klst)/2)