Use Laplace Approximations to improve linear approximations to the posterior

Description

The approximation provided by linear.bayes can be improved by performing Laplace approximations. This function is a development version of a wrapper to do that for all of the returned by linear.bayes.

Usage

adjust.linear.bayes(lbo, ana.obj=lbo$call$ana.obj, ...)


Arguments

 lbo The object returned by linear.bayes ana.obj The analysis.object used to create lbo. This need not be given explicitly, iff the original version is in the search path. ... currently unused

Value

A list of class "adjust.linear.bayes" containing:

 odds A vector, typically of length k giving the odds for models of size 1, 2, ..., k under a uniform posterior relative to a model with no genes. loc.posterior The marginal posterior probabilities by locus coefficients The marginal posterior means of the coefficients one.gene.adj Results of fits for one gene models n.gene.adj Results of fits for modles with more than one gene call the call to adjust.linear.bayes

Note

For large linear.bayes objects invloving many gene models, this can require a very long time to run.

Author(s)

Charles C. Berry cberry@ucsd.edu

linear.bayes