You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I was almost going to raise a new issue, and then I found this. So here's more on why the above is important:
Summary:
Sometimes we want to fit a PK/PD model with fixed PK parameters. Currently however, users must pass both PK and PD parameters in the array theta. We should relax this, and allow users to pass variables in the ODE system using x_r and x_i (i.e. real and integer data). This should lead to an appreciable speed-up.
Even more so, in fact, with the mixed solver (for the Friberg-Karlsson model, the theoretical speed-up would be around 70 - 80 %). The coding shouldn't be too bad; we need to make sure we promote variables only when necessary.
The only major drawback I see (other than development effort) is that this would make the function signature heavier. It might be worth thinking about ways around it.
Summary:
In Pred1_general_solver.hpp, pred1 calls pmetrics_solver as:
Right now, all the parameters are expected to be stored in theta (which is fine), but there is no mechanism for passing dat or dat_int.
Current Version:
Stan v2.14.0
Torsten v0.82
The text was updated successfully, but these errors were encountered: