|
Abstract:
|
The UNAIDS has decided to use Bayesian melding as the basis for its probabilistic HIV prevalence projections. The Bayesian melding approach combines expert opinion, past trends in inputs and outputs, and it provides a framework for assessing uncertainty in mechanistic models in general. The posterior distribution was produced by sampling-importance-resampling procedure. For some countries, however, it is not efficient to give a good picture of the uncertainty because the posterior distribution tends to be multimodal and to feature nonlinear sharp ridges. In this article, we present a more efficient Incremental Mixture Importance Sampling (IMIS), while retaining its essential simplicity and transparency, and so can be used for all the countries involved. It also leads to a simple estimator of the integrated likelihood that is the basis for Bayesian model comparison and model averaging.
|
