Neural network approximations to posterior densities: an analytical approach
Year: 2003
Type: preprint
Abstract: In Hoogerheide, Kaashoek and Van Dijk (2002) the class of neural network\nsampling methods is introduced to sample from a target (posterior)\ndistribution that may be multi-modal or skew, or exhibit strong correlation\namong the parameters. In these methods the neural network is used as an\nimportance function in IS or as a candidate density in MH. In this note we\nsuggest an analytical approach to estimate the moments of a certain (target)\ndistribution, where `analytical' refers to the fact that no sampling\nalgorithm like MH or IS is needed.We show an example in which our analytical\napproach is feasible, even in a case where a `standard' Gibbs approach would\nfail or be extremely slow.
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W1587285391