4 ms·
+ Monte Carlo integration goes to shit in higher dimensions...
by physicsguy 6y ago
+ Monte Carlo integration goes to shit in higher dimensions...
- aaronchall 6y agoThat's why we do Hamiltonian Monte Carlo in higher dimensions.
- srean 6y agoBeats the trapezoid rule hands down though, in fact that was the main selling point of MC over other quadrature methods. MC error goes down as O 1/sqrt(n) with quasi MC one can get it down to O 1/n.
- teorema 6y ago+ MCMC can be set up in a very model-agnostic way. Analytic methods are really useful when they can be done, but often have to be worked out for each specific model.
- jumpinggenes 6y agoGibbs sampling to the rescue.