$^1$Berkeley, $^2$Harvard
univariate — an object (function, model, etc) that is dependent on one variable. In contrast to multivariate, which is an object that is dependent on the specification of multiple variables.
Gibbs sampling — an Markov chain Monte Carlo algorithm used to sample a multivariate probability distribution. It is a special case of the more general Metropolis-Hastings algorithm and samples each variable of a multivariate model separately, but not indepdently, as each sample is dependent on the value of the other variables. Very lengthy Wikipedia entry.
This paper can be found in its entirety on Project Euclid. Views expressed here are my own and do not represent those of my institution or my collaborators.
In [ ]: