search

bonStats / tidytreatment

Tidy methods for Bayesian treatment effect models
Other
14 stars 2 forks source link

Great package! #10

Open schelhorn opened 2 years ago

schelhorn commented 2 years ago

Hey, very nice package! Are you continuing the development? Please do so, it looks super useful already! Perhaps adding bartCause would be an option as well?

bonStats commented 2 years ago

Hi @schelhorn, yes it's still being developed. I will look into adding bartCause. Thank you for bringing it to my attention!

schelhorn commented 2 years ago

Sweet, thank you. There's also an associated package for making plots from bartCause models, if you are interested, as well as a Shiny app.

Another topic of interest of myself (and I believe also other pharmaceutical researchers) is subgroup analysis. I believe it is possible to use the tree structure of BARTs to estimate subgroups of subjects with heterogenous treatment effects, and perhaps even identify the covariates causally associated with the latter. Would that be something that you could add a short section in the {tidytreatment} vignette about?

bonStats commented 1 year ago

{bartCause} supported since https://github.com/bonStats/tidytreatment/tree/e740bcad6659fcbab52a16e1b809374ee4f4d8ba

I will keep this issue open to remind me about the subgroup/plotting aspects you have raised.