Closed dajmcdon closed 2 months ago
@simonpcouch Thanks!
Can we call it "quantile_pred"
? We have a special "class_pred"
vctrs class in probably.
Thanks for making this
Or we can revisit if you'd really prefer the tibble list-col version.
Nope, I'm good with using this class!
One other small thing... let's remove row and column names from values
in parsnip_quantiles()
. They carry over into the vctrs object (although you can't see them unless you do something like unclass()
)
@dajmcdon Merge time!
One note... after looking into performance metrics, we realized that quantile_pred()
and its methods will also be needed in yardstick. We want to avoid adding parsnip as a dependency there.
For these cases, we move infrastructure to hardhat (which everyone depends on at some level). After I merge this, I’ll PR into hardhat to put the functions there, and then PR to parsnip to remove them.
This pull request has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue.
@topepo
quantile_levels
).rq()
predictions (1 or moretau
/quantile_levels
).I added a test for the constructor and adjusted your tests for the
quantreg
engine.Don't hesitate to ask for more changes! Or we can revisit if you'd really prefer the tibble list-col version.