-
Notifications
You must be signed in to change notification settings - Fork 12
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Develop #10
Conversation
pat-alt
commented
Oct 15, 2022
- Wrapping models not machines.
- Distinction between inductive and transductive.
- Added Jackknife.
…result, Xnew) on wrapped EvoTrees models fails for some reason.
…ctions for classifiers, I need access to data -> labels(y). Not sure how to do this when treating ConformalModels just as Model, since only machines have access to data.
5 wrap models not machines
|
||
# Conformal prediction through dispatch: | ||
""" | ||
MMI.predict(conf_model::ConformalModel, Xnew, coverage::AbstractFloat=0.95) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Type piracy? MMI.predict(::Model, ...)
will always have a fitresult
as the second argument here. And your method has an extra parameter coverage
, which doesn't really fit into the MLJ API at all, unless you make it a hyperparameter (field of the model).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yep. We have already have Interval <: Supervised
but maybe can extend that further. Related is helping me out with this list of "target proxies" (predictions types) for the future.