Closed djhocking closed 9 years ago
Hmm, well I'm not exactly sure what your asking. But yeah, the web server can save output plots to pdf, which you can then access after the run is complete. Probably a good idea to leave this diagnostic-type output in the scripts for now so we know if its working properly. Later we can add a DEBUG flag that will tell the script whether to save diagnostic output or not. Get to that later once we get the whole thing running.
Sounds good, I just wanted to make sure that the way ggmcmc was saving plots wasn't going to mess something up. I like the idea of adding a DEBUG flag later.
I have currently been using
ggmcmc
(package and function) to make and save the traceplots from the model to ensure mixing and model convergence. I'm not sure if this will work with the web system and running the bash scripts or not. I suspect it will be okay but I'm not sure. I'm also not really sure if it's needed every time the model is run.As long as the mcmc.list object is saved (M.ar1) I could always do it in a separate R script outside of the web system on my local machine if I wanted to check.
Example: