Incremental Gibbs sampling

From Cohen Courses
Revision as of 00:05, 5 April 2011 by Yandongl (talk | contribs) (Created page with 'Wikipedia page on Gibbs sampling. Usually Gibbs sampling works in a batch mode. By generating sequence of Markov chain, the stationary state of the chain is the final distributi…')
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigationJump to search

Wikipedia page on Gibbs sampling.

Usually Gibbs sampling works in a batch mode. By generating sequence of Markov chain, the stationary state of the chain is the final distribution we want. Incremental Gibbs sampling, specifically for LDA, can start with some initial states, and continue the sampling process for new docuemnts

Retrieved from "http://curtis.ml.cmu.edu/w/courses/index.php?title=Incremental_Gibbs_sampling&oldid=5525"