more complex bootstrap summarization methods
Opened this issue · 2 comments
Any plans on integrating more complex bootstrap analyses (e.g. different CI methods) into the package?
I have similar functions in rsample and adding more post-bootstrap analysis techniques is on my to-do list. There's probably no point in duplicating effort.
My plan was to port code from boot or bootstrap and/or make conversion functions that will take my objects and make them compatible with the objects from one of those packages.
That's a good idea – you can pass tidyboot() any functions to compute over the samples, so it would just be a matter of adapting other CI methods to functions that operate over tidy dfs, which should be pretty straightforward. Those functions might be reusable across our packages?
I've looked at functions in these two packages and, for rsample at least, some CI methods are easy to adapt and others aren't. Some require the original statistic generating function etc.