stanfordmlgroup / ngboost

Natural Gradient Boosting for Probabilistic Prediction
Apache License 2.0
1.65k stars 215 forks source link

Any plan for model explanation functionality? #5

Closed zhiruiwang closed 4 years ago

zhiruiwang commented 5 years ago

Read the paper and tried this package, had to say it is marvelous! I will not hesitate to use it in my real-life work, but wonder if there is plan to add support for model explanation tool such as feature importance plot, SHAP plot or tree visualizer. It would be crucial to present the model explanation to business stakeholders. Thanks for this fantastic masterpiece!

avati commented 5 years ago

Hi @zhiruiwang, ngboost is still in very early stage of development. We are working on several features and speedup changes. It is very likely that something like feature importance might be added down the line. We also welcome contributions :-)

zhiruiwang commented 4 years ago

@avati I have made a pull request #27 of feature importance plot for tree learner using similar API as sklearn. I will also make a pull request to add in SHAP support for NGBoost.

zhiruiwang commented 4 years ago

SHAP support added to the same pull request #27

alejandroschuler commented 4 years ago

https://github.com/stanfordmlgroup/ngboost/pull/57 https://github.com/stanfordmlgroup/ngboost/pull/58