-
-
Notifications
You must be signed in to change notification settings - Fork 1k
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
Modelling issues #2125
Comments
@pavlin-policar could you take a look? |
Yes, sure thing. I'll get right on it. As far as 4. is concerned, I don't know what would make more sense, because the |
I think this has the lowest priority and can wait for some discussion and consensus. I agree it makes sense in regression where it contains multiple classes. If it will remain with only one class in modelling, I think it makes less sense. But it is only a name after all, so several choices are good enough. |
@lanzagar Can this be considered resolved? |
Sure, why not. |
Orange version
3.4.1
Issues
Orange/__init__.py
and needs to be explicitly imported (unlike other submodules)Orange/modelling/__init__.py
(i.e. Constant and SGD)__all__
so things are pulled intomodelling
that should not belinear.py
submodule, but if it will only contain SGD it could be renamed to e.g. sgd.py?_change_kwargs
. I would like/expect defaults to be used when not providing specific losses etc.Learner.__call__
instead ofSklLearner.__call__
so the attributeused_vals
is not set, andparams
does not exist. This is a problem at least in [FIX] Widget Logistic Regression: can handle unused values #2116, but could have other implications. I think probability predictions are not expanded correctly (seeSklModelClassification.__call__
or compare with e.g. logreg on first half of iris)!The text was updated successfully, but these errors were encountered: