Skip to content

Suggestion: implement SHAP for model interpretability #10

@lukethehuman

Description

@lukethehuman

SHAP (SHapley Additive exPlanations) is a game theoretic approach to explain the output of any machine learning model. Suggest exploring where SHAP may have application for SLEDO.

Resources:
https://shap.readthedocs.io/en/latest/
https://www.nature.com/articles/s42256-019-0138-9
https://papers.nips.cc/paper_files/paper/2017/hash/8a20a8621978632d76c43dfd28b67767-Abstract.html
https://towardsdatascience.com/interpretable-machine-learning-with-xgboost-9ec80d148d27

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions