1Blurb::
2Calculate model evidence (marginal likelihood of model) when using Bayesian methods
3
4Description::
5Model evidence is used in Bayesian model selection and model averaging.
6It is defined as the probability of the data given the model, and
7is calculated by averaging the likelihood of the model parameters
8over all values of the model parameters according to their prior
9distributions. In Dakota, one must calculate the
10model evidence separately for each model and perform the normalization
11to obtain the posterior model plausibility for each model.
12
13<b> Default Behavior </b>
14
15When specifying \c model_evidence, there are two methods of
16calculating it.  One or both may be specified.  They
17include the Monte Carlo approximation, given by \c mc_approx
18and the Laplace approximation, given by \c laplace_approx.  \c mc_approx
19is the default approach.
20
21<b> Expected Output </b>
22Currently, the model evidence will be printed in the screen output
23with prefacing text indicating if it is calculated by
24Monte Carlo sampling or the Laplace approximation.
25
26<b> Usage Tips </b>
27
28
29Topics::
30
31Examples::
32
33Theory::
34Faq::
35See_Also::
36