One of the problems of going to a conference called “MCMSki” is that it is hard to persuade people that you are working rather than extending your Christmas break. These pictures will probably not help. Continue reading

# Author Archives: Martyn

# MCMSki 2014

I am off to the MCMSki 2014 meeting in Chamonix. I am talking twice. The first talk is in a session on *Computational and Methodological Challenges in evidence synthesis and multi-step (modular models). *The title is a bit of a mouthful, but I will be talking about cuts. There is a demand for a cut function in JAGS but it is incredibly hard to implement it properly. In fact the cut function, as currently implemented in OpenBUGS does not work correctly. I have been saying this for years, without making any real progress or providing a solution in JAGS. I agreed to give the talk in order to explain myself to an expert audience.

The second talk is in a round-table discussion on *Recent Developments in Software for MCMC** *with other software developers Andrew Thomas (BUGS), Bob Carpenter (Stan) and Adrien Todeschini (BiiPS). BiiPS is a fork of JAGS that uses the same front end, but sequential Monte Carlo as a back end instead of MCMC.

I will report back after the meeting. In the mean time, you should be able to find a conference diary by Christian Robert at Xi’an’s Og.

# Postdoctoral Opportunity at IARC

My colleague Pietro Ferrari has funding for a postdoctoral position in biostatistics in the Nutritional Epidemiology Group of the International Agency for Research on Cancer in Lyon, France. This is part of a project entitled “Integrated approach for dietary exposure and biomarker measurements in aetiological models”. I am a co-investigator on this project. See below the fold for more details. Continue reading

# BayesComp web site

Last year the International Society for Bayesian Analysis launched a new section on Bayesian Computing. The BayesComp section now has its own “website of information, tools, and advice as an authoritative central resource” at bayescomp.wikidot.com.

If you are not already a member of BayesComp then you should definitely think about joining. See the ISBA web site for more details.

# Rencontres R, Lyon 27-28 June

Last year, the first French-speaking R conference, “Rencontres R” was held in Bordeaux. The meeting was a great success, and a second one will be held in Lyon on 27 and 28 June 2013.

The abstract submission deadline of 7 April is fast approaching. Three types of presentation are planned for the meeting:

- 20 min regular talks,
- 6 min Lightning Talks,
- Posters.

To submit an abstract, follow the instructions on the conference web site.

# Course: Statistical Practice in Epidemiology with R

*Statistical Practice in Epidemiology with R* is a one week course aimed at promoting the use of R among epidemiologists. This annual course has been running almost every year since 2004. This year, the course will take place from 23 to 29 May 2012 at the International Agency for Research on Cancer, in Lyon, France.

For more details about the course, including how to apply, see the SPE home page. The deadline for applications is April 1 2013.

# A tutorial on extending JAGS

The JAGS-Wiener module by Dominik Wabersich and Joachim Vandekerckhove extends JAGS by adding the distribution “dweiner” for the the first passage time of a drift diffusion process. As far as I know, this is the first third-party JAGS extension that is distributed as a standalone module [1]. You can download the source code from sourceforge, and for Windows users, there is a binary installer.

Dominik and Joachim have documented the process of putting together the JAGS-Wiener module in a tutorial article, which is now submitted to Behavior Research Methods. This article fills a much needed gap in the documentation, and will hopefully encourage others interested in extending JAGS to follow suit.

[1] There is also the lossDev package for R, distributed via CRAN, which also includes JAGS extensions. This is a little different as it is tied to the R interface.

# Bob O’Hara on Nate Silver’s election predictions

Bob O’Hara has written a nice article for The Guardian explaining the statistical techniques underlying Nate Silver’s remarkably accurate predictions of the US presidential election. Bayes’ Theorem and hierarchical modelling both get a name check, but there are no equations. The article also includes a selection of #natesilverfacts.

# ggmcmc – diagnostic plots for MCMC with ggplot2

Xavier Fernández i Marín, who maintains the jags package on Gentoo Linux, writes to tell me he is developing the R package ggmcmc. This package is for visualizing Markov Chain Monte Carlo output using ggplot2 graphics and should complement the existing plots for base and lattice graphics provided by coda. A comparison of all three graphical styles is given below. Continue reading

# JAGS 3.3.0 for Mac OS X

Bill Northcott’s binary distribution of JAGS 3.3.0 for Mac OS X is now available. This distribution supports 10.8 (Mountain Lion) and 10.7 (Lion). As explained by Bill in the installation manual, changes in Apple’s developer tools make it difficult to support earlier versions.