Design Ideas for Markov Chain Monte Carlo Software
نویسنده
چکیده
This article discusses design ideas useful in the development of Markov chain Monte Carlo (MCMC) software. Goals of the design are to facilitate analysis of as many statistical models as possible, and to enable users to experiment with different MCMC algorithms as a research tool. These ideas have been used in YADAS, a system written in the Java language, but are also applicable in other object-oriented languages.
منابع مشابه
A Framework for Expressing and Estimating Arbitrary Statistical Models Using Markov Chain Monte Carlo
YADAS is a new open source software system for statistical analysis using Markov chain Monte Carlo. It is written with the goal of being extensible enough to handle any new statistical model and versatile enough to allow experimentation with different sampling algorithms. In this paper we discuss the design of YADAS and illustrate its power through five challenging examples that feature unusual...
متن کاملMarkov Chain Monte Carlo Methods for Bayesian Data Analysis in Astronomy
Markov Chain Monte Carlo based Bayesian data analysis has now become the method of choice for analyzing and interpreting data in almost all disciplines of science. In astronomy, over the last decade, we have also seen a steady increase in the number of papers that employ Monte Carlo based Bayesian analysis. New, efficient Monte Carlo based methods are continuously being developed and explored. ...
متن کاملEmbarrassingly parallel sequential Markov-chain Monte Carlo for large sets of time series
Bayesian computation crucially relies on Markov chain Monte Carlo (MCMC) algorithms. In the case of massive data sets, running the Metropolis-Hastings sampler to draw from the posterior distribution becomes prohibitive due to the large number of likelihood terms that need to be calculated at each iteration. In order to perform Bayesian inference for a large set of time series, we consider an al...
متن کاملMachine Learning, Information Retrieval, and Record Linkage
Classification into groups using terms available in the data underlies machine learning, information retrieval, and record linkage. Classifiers such as Bayesian networks in machine learning and term weighting in information retrieval depend primarily on training data sets for which truth is known. These classifiers may be relatively slow to adapt to new situations in which new data have charact...
متن کامل