interpretable-ml / iml

Interpretable ML package designed to explain any machine learning model.
60 stars 7 forks source link

Feature labels in AdditiveForceVisualzer #2

Closed ytsaig closed 6 years ago

ytsaig commented 6 years ago

Thanks for putting together SHAP and the associated visualization tools, really well done!

Question: would it be easy to add tooltips with the feature names (and associated values) when hovering over individual elements in the single AdditiveForceVisualizer? Right now the labels for the top few features are displayed below the bars, but with many features it'd be useful to know what are other features that contribute to an individual prediction. My javascript is pretty weak, so can't tell if it's easy to do.

slundberg commented 6 years ago

great idea! I'll leave this issue open as a tag to do that. It should not be very hard.

On Tue, Apr 10, 2018 at 4:52 PM ytsaig notifications@github.com wrote:

Thanks for putting together SHAP and the associated visualization tools, really well done!

Question: would it be easy to add tooltips with the feature names when hovering over individual elements in the single AdditiveForceVisualizer? Right now the labels for the top few features are displayed below the bars, but with many features it'd be useful to know what are other features that contribute to an individual prediction. My javascript is pretty weak, so can't tell if it's easy to do.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/interpretable-ml/iml/issues/2, or mute the thread https://github.com/notifications/unsubscribe-auth/ADkTxTSB9Na_pD0-GmdV5ntrPJ9icaXpks5tnUXOgaJpZM4TPKwT .