Bayes Factors and Multimodel Inference

摘要: Multimodel inference has two main themes: model selection, and averaging. Model averaging is a means of making conditional on set, rather than selected model, allowing formal recognition the uncertainty associated with choice. The Bayesian paradigm provides natural framework for averaging, context evaluation commonly used AIC weights. We review multimodel inference, noting importance Bayes factors. Noting sensitivity factors to choice priors parameters, we define propose nonpreferential as offering reasonable standard objective inference.