Metropolis Paclet SymbolMore About »

BlockMetropolis

BlockMetropolis[logpdf, logpdfjump, randomjump, x0, n]
generates n updates for a set of conditional independent random variables starting from x0. The target distribution is p[x] with the proposal distribution J[xt, x] randomized by x = randomjump[xt] given current xt. For any subset x[[indices]], logpdf[x, indices] = Log[p[x[[indices]]]] and logpdfjump[xt,x, indices] = Log[J[xt[[indices]],x[[indices]]]].
  • To use BlockMetropolis, you first need to load the Metropolis package using Needs["Metropolis`"].
  • The BlockMetropolis has same options of Metropolis.
  • Each update is given by {x, logpdf[x], 1/0 (acccept/not), Min[1,Metropolis-Hastings ratio]}.