vdorie / bartCause

Causal Inference using Bayesian Additive Regression Trees
77 stars 10 forks source link

Request for vignette #5

Open meghapsimatrix opened 4 years ago

meghapsimatrix commented 4 years ago

Hi, Could I request a vignette that goes over examples on how to use bartc with applied data? How to deal with missing data, multilevel structure etc and any other things to look out for when using bartc?

schelhorn commented 2 years ago

I support this request; a vignette would be really useful for people starting to work with the package.

It would also be awesome if the vignette could cover feature importances for the features that are specific for the treatment effect (i.e., where the features are predictive of the responses for treated subjects only, and not only prognostic [prognostic features affect the response regardless of treatment status]).