[MRG] Multi class label documentation #2207
Conversation
| Multilabel learning with OvR | ||
| ---------------------------- | ||
| Multilabel learning | ||
| ------------------- | ||
|
|
||
| :class:`OneVsRestClassifier` also supports multilabel classification. | ||
| To use this feature, feed the classifier a list of tuples containing |
There was a problem hiding this comment.
You can use also a binary indicator format.
I think that a small example would be nice which would basically show how to use that estimator with either of list of tuple or binary indicator format.
|
Could you add a pointer to LabelBinarizer? |
| @@ -7,23 +7,34 @@ Multiclass and multilabel algorithms | |||
|
|
|||
| .. currentmodule:: sklearn.multiclass | |||
|
|
|||
| This module implements multiclass and multilabel learning algorithms: | |||
There was a problem hiding this comment.
I think it say somewhere very high up that you don't need this meta-estimators to do multi-class classification as all classifiers have build-in multiclass support
There was a problem hiding this comment.
Most estimators, I believe.
There was a problem hiding this comment.
all. without exception. if there is an exception, it should be listed below as not supporting multiclass imho.
|
@schwarty do you intend to finish this before the release? |
|
@schwarty : we are planning to release during the week end. It would be good if you could address the comments. |
|
please push so we can review your changes :) |
|
It should be (almost) ready to merge. |
|
@schwarty : you should push your changes: we want to merge this guy :) |
|
Close in favor of #2269. |
Re-organized the documentation, added some real simple examples for the definitions of multiclass and multilabel classification.