Is your feature request related to a problem? Please describe.
Example nb for multi label classification uses the standard interp.plot_confusion_matrix() is used, but this has no meaning. The proper confusion matrices are computed lower but not as a plot. To make it more readable for newbies, a set of 2 x 2 plots would be helpful
Describe the solution you'd like
Add a function taking the interp object, and uses sklearn multilabel_confusion_matrix() to get each individual confmatrix and plot then in a grid to make it visual.
This will allow to make the example nb more readable for newbies and also can be reused for other
Describe alternatives you've considered
We also can create a class MultiLabelClassificationInterpretation, that inherits from ClassificationInterpretation and adds the method .plot_multilabel_confusion_matrices()
Additional context
Code for the function is ready. Just need to decide if we do as a function or an Class method. Guidance on preferred option and where to add it appreciated.
Is your feature request related to a problem? Please describe.
Example nb for multi label classification uses the standard
interp.plot_confusion_matrix()is used, but this has no meaning. The proper confusion matrices are computed lower but not as a plot. To make it more readable for newbies, a set of 2 x 2 plots would be helpfulDescribe the solution you'd like
Add a function taking the
interpobject, and usessklearn multilabel_confusion_matrix()to get each individual confmatrix and plot then in a grid to make it visual.This will allow to make the example nb more readable for newbies and also can be reused for other
Describe alternatives you've considered
We also can create a class
MultiLabelClassificationInterpretation, that inherits fromClassificationInterpretationand adds the method.plot_multilabel_confusion_matrices()Additional context
Code for the function is ready. Just need to decide if we do as a function or an Class method. Guidance on preferred option and where to add it appreciated.