Establishing some order amongst exact approximation MCMCs
Duration: 30 mins 23 secs
Share this media item:
Embed this media item:
Embed this media item:
About this item
Description: |
Vihola, MS (University of Jyväskylä)
Wednesday 23 April 2014, 11:05-11:40 |
---|
Created: | 2014-04-28 16:52 |
---|---|
Collection: | Advanced Monte Carlo Methods for Complex Inference Problems |
Publisher: | Isaac Newton Institute |
Copyright: | Vihola, MS |
Language: | eng (English) |
Abstract: | Co-author: Christophe Andrieu (University of Bristol)
Exact approximation Markov chain Monte Carlo (MCMC) algorithms are a general class of algorithms for Bayesian inference in complex models. We discover a general sufficient condition which allows to order two implementations of such algorithms in terms of mean acceptance probability and asymptotic variance. The key condition is convex order between the weight distributions, which emerges naturally when the weight distributins stem from importance sampling approximations with different number of samples. |
---|
Available Formats
Format | Quality | Bitrate | Size | |||
---|---|---|---|---|---|---|
MPEG-4 Video | 640x360 | 1.94 Mbits/sec | 442.50 MB | View | Download | |
WebM | 640x360 | 602.34 kbits/sec | 134.12 MB | View | Download | |
iPod Video | 480x270 | 521.67 kbits/sec | 116.09 MB | View | Download | |
MP3 | 44100 Hz | 249.76 kbits/sec | 55.64 MB | Listen | Download | |
Auto * | (Allows browser to choose a format it supports) |