generalized random forests #1299
Conversation
hfrick
left a comment
hfrick
left a comment
There was a problem hiding this comment.
Engine-specific tuning parameters are not part of this (yet?) - that's on purpose, right?
I've focused on the changes directly related to grf, the others looked mostly like air-related.
|
|
||
|
|
||
| # ------------------------------------------------------------------------------ | ||
| # grf components |
There was a problem hiding this comment.
Nit: It'd be nice if the ordering here was consistent. Either each "thing" for all modes or everything for one mode, then the next mode.
There was a problem hiding this comment.
If I understand correctly, the pattern that is in parnsip is to run all of the set_model_engine() lines for each mode then all of the set_dependency() calls for each mode.
I don't think that we are ordering anything different for grf than the others in this file or in other files.
| #' The \pkg{grf} fits models that create a large number of decision | ||
| #' trees, each independent of the others. The final prediction uses all | ||
| #' predictions from the individual trees and combines them. |
There was a problem hiding this comment.
This reads very general, and usually we try to be a bit more specific in this description block. Is there anything else we could say here? I do realize that this is about generalized random forests so there might not be!
There was a problem hiding this comment.
I made a very small tweak, but I don't really think that there is anything additional to say. It's a generalization of random forests (and not different from the other details files). We do have a reference, so people can look there for more details.
The main reason we have that description block is to avoid saying "random forest" since some people may not know what it is. It's not meant to be very detailed.
Adds engines for classification, regression, and quantile regression.
Tests in tidymodels/extratests#281