-
Notifications
You must be signed in to change notification settings - Fork 22
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
feat(_find_ml_task): Refine ML task detection logic #1305
Conversation
da17821
to
76dc0fd
Compare
Coverage Report for backend
|
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 is also another "bug" that I realised. In l.99-100, we return "unknown"
.
In short, we use the estimator and if the estimator does not provide us the require information then we return "unknown". I think it would be better to fall back to the target inference, and try to get the ML task from it.
I realized that if someone provide a non-compatible estimator, we would not be inferring the task properly.
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.
For the rest of the PR, all look good.
e7a7a96
to
ba2168c
Compare
ML detection now also works for multioutput targets (e.g. where
y
is a 2d array).Added 3 new MLTask variants:
multioutput-binary-classification
,multioutput-multiclass-classification
andmultioutput-regression
.regression
still means "single-output regression".The detection uses the same mechanism described in be29a26 to discriminate between an array of integers that is actually for classification vs. one that is for regression.
Closes #1005