** Control structures: if-then-else, for and while loops ** Permits sub-models with very different stochastic hierarchies ** Large selection of built-in probability distributions In my workshop I have a slide headed “Why Stan?” with the following response: Īs you say the primary motivations for using Stan for pharmacometric applications fall under the headings of flexibility and the effectiveness/efficiency of the sampler. This entry was posted in Bayesian Statistics, Statistical computing and tagged ODE, pharmacology, PK/PD by Bob Carpenter. There’s more in the works, but the above are the top of our to-do list. Then we could replace the max marginal likelihood approach of NONMEM with a very speedy variational inference mechanism allowing much more general models. We also need to evaluate how well variational inference works for ODE problems. And of course, by Michael Betancourt working out all the math and Daniel, Michael, and I working out the code with Sebastian’s and Wenping’s input. These new designs are largely being guided by Sebastian Weber and Wenping Wang at Novartis. Next, hopefully by Stan 2.10, we’ll have a stiff solver and maybe a way for users to supply analytic coupled-system gradients and Jacobians. We should also allow user-defined control of absolute and relative tolerances. And we’ve really sped up the Jacobian calculations when Michael Betancourt realized we were doing a lot of redundant calculation and he and I put a patch in to fix it. The next minor release of Stan (2.9) should stop the freezing issue when parameters wander into regions of parameter space that lead to stiff ODEs. There’s been lots of behind-the-scenes activity on our ODE solvers-we’re really just getting burned in warmed up. Maybe Bill will jump in with some other motivations. Or a good NONMEM-like event data language on top. So far, though, we’re in the hole in not having a stiff ODE solver in place. In case you’re wondering why people would use Stan for this instead of something more specialized like Monolix or NONMEM, it’s because of the modeling flexiblity provided by the Stan language and the effectiveness of NUTS for MCMC. Having said that, we’ve learned a lot from Bill and colleagues on our mailing lists as we were designing ODE solvers for Stan (an ongoing issue-see below for future plans).īill’s tutorial is up against a 2-day Monolix tutorial and a 2-day tutorial on R by Devin Pastoor, who’s also been active on our mailing lists recently. This is super cool for us, because Bill’s not one of our core developers and has created this tutorial without the core development team’s help. Thursday 8 October 2015, 8 AM - 5 PM, Crystal City, VA Getting Started with Bayesian PK/PD Modeling Using Stan: Practical use of Stan and R for PK/PD applications.Bill Gillespie, of Metrum, is giving a tutorial next week at ACoP:
0 kommentar(er)
