Use iterated Laplace approximation as a proposal for importance
 sampling or the independence Metropolis Hastings algorithm.

 misc 

Importance Sampling and independence Metropolis Hastings sampling

The iterLap (iterated Laplace approximation) algorithm approximates a
general (possibly non-normalized) probability density on R^p, by repeated
Laplace approximations to the difference between current approximation
and true density (on log scale). The final approximation is a mixture of
multivariate normal distributions and might be used for example as a
proposal distribution for importance sampling (eg in Bayesian applications).
The algorithm can be seen as a computational generalization of the Laplace
approximation suitable for skew or multimodal densities.

Bjoern Bornkamp

iterLap

Approximate Probability Densities by Iterated Laplace
Approximations

Importance Sampling and independence Metropolis Hastings sampling function

<dl> <dt>obj</dt>
<dd>an object of class "mixDist"</dd> <dt>nSim</dt>
<dd>number of simulations</dd> <dt>df</dt>
<dd>degrees of freedom of the mixture of t distributions proposal</dd> <dt>post</dt>
<dd>log-posterior density</dd> 
 <dt>vectorized</dt>
<dd>Logical determining, whether <code>post</code> is vectorized</dd> <dt>cores</dt>
<dd>number of cores you want to use to evaluate the target density
 (uses the mclapply function from the parallel package). For
 Windows machines, a value &gt; 1 will have no effect, see mclapply
 help for details.</dd> <dt>...</dt>
<dd>additional arguments passed to <code>post</code>.</dd></dl>

Arguments

Author

Monte Carlo sampling using the iterated Laplace approximation. — Importance Sampling and independence Metropolis Hastings sampling

<dl>

 <dt>obj</dt>
<dd>an object of class "mixDist"</dd>

 <dt>nSim</dt>
<dd>number of simulations</dd>

 <dt>df</dt>
<dd>degrees of freedom of the mixture of t distributions proposal</dd>

 <dt>post</dt>
<dd>log-posterior density</dd>

 
 <dt>vectorized</dt>
<dd>Logical determining, whether <code>post</code> is vectorized</dd>

 <dt>cores</dt>
<dd>number of cores you want to use to evaluate the target density
 (uses the mclapply function from the parallel package). For
 Windows machines, a value &gt; 1 will have no effect, see mclapply
 help for details.</dd>

 <dt>...</dt>
<dd>additional arguments passed to <code>post</code>.</dd>

</dl>

Importance Sampling and independence Metropolis Hastings sampling: Monte Carlo sampling using the iterated Laplace approximation.

Description

Usage

Value

Arguments

Author

Examples