This has the following advantages:
In the event of a power outage or other unforeseen system crash (eg, Windows 10 update), you still have most of the results so far.
You can get interim results without stopping the main process: start a new instance of R and read in the files generated so far. Abort the run if already good enough.
No need to exactly calculate the time needed for a run: make sure you ask for enough iterations, then abort when time's up.
Reduce memory needed in R (and never again get a "cannot allocate..." error after a long JAGS run): save as many iterations and parameters as you like to disk, then load only a subset of parameters or thin the chains before loading into R.
If some chains exit with errors (usually due to incompatible starting values), you can still recover the results for those that ran properly.
For more information see here.
mcmcOutput
packageA new - and still being developed - package, mcmcOutput
, provides a class for storing MCMC output and a range of methods and functions to manipulate, summarise and display output. More information is here. You need to install mcmcOutput
from GitHub before installing saveJAGS
.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.