Bayesian Methods and Modeling
A recent article in Science magazine (286, 19 November 1999, pp. 1460–1464) noted the rejuvenation of the Bayesian inference paradigm and its increasingly warmer reception in the scientific community. Pertinent here, is its recent application to a wide range of environmetrics problems, allowing inferential capability that may not be available using classical methodology. This article provides a brief overview of Bayesian methodology and modeling beginning with Bayes' theorem. It then elaborates the Bayesian inference paradigm and the Bayesian modeling perspective. Historically, computation was the major impediment to the use of Bayesian analysis for complex models. Now, the wide availability of inexpensive, fast computing combined with breakthroughs in simulation-based model fitting have removed this impediment. Bayesian models, which were formerly infeasible to examine, can now be fitted. More realistic models can be investigated.
