iterLap: Approximate Probability Densities by Iterated Laplace
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.
Please use the canonical form
to link to this page.