Sequentially Adaptive Bayesian Learning

Lead CI: John Geweke

The SABL algorithm is a framework for integration (in particular, and with emphasis on, Bayesian inference) and for optimization (in particular, and with emphasis on, M-estimation). It is a generalization of existing sequential Monte Carlo methods that also incorporates ideas and procedures from sequential Monte Carlo, evolutionary algorithms, simulated annealing, and Markov chain Monte Carlo. There are quite a few specific variants of the algorithm: some well-established, others developed by the project within the last year, and more to come in the near future. The project also includes a variety of statistical models and provides a platform on which others can incorporate other models easily.