Managing simulation results with rstansim

knitr::opts_chunk$set(collapse = T)

Once data has been simulated and models fit the next stage is to organise the results for analyse/storage and then to extract the relevant data.

Managing multiple simulation objects

Any simulation study is likely to involve multiple stan models, and therefore likely to involve multiple calls to fit_models() and multiple output objects from these calls. Held seperately these increase the complexity of analysing the simulation results. For this reason, in cases where multiple simulation objects are being managed it is strongly recommended that the collect() function is used to group them into a single object of class stansim_collection.

See the end to end example for an example of how to use the collect() function to combine multiple stansim_simulation objects in to one. collect() can be used to combine both stansim_simulation and stansim_collection objects together. Once all results for a single simulation study have been combined further analysis and storage only needs to focus on a single R object.

Extracting relevant data

Once data is collected into a single object it is then best extracted using the packages built in functions, all prefixed with extract_. These functions currently include:

Any other stored data can be extracted manually, but this process can be quite involved due to the deep structure of the stansim_simulation and stansim_collection objects.

Renaming objects

In order that all simulation data is clear and understandable it's important that informative object names are set. If for whatever reason they're not set when initially constructed both stansim_simulation and stansim_collection objects can have their names reset by the rename() function.



Try the rstansim package in your browser

Any scripts or data that you put into this service are public.

rstansim documentation built on Sept. 22, 2017, 1:06 a.m.