Practical guide to SHAP analysis: Explaining supervised machine learning model predictions in drug development

Abstract Despite increasing interest in using Artificial Intelligence (AI) and Machine Learning (ML) models for drug development, effectively interpreting their predictions remains a challenge, which limits their impact on clinical decisions. We address this issue by providing a practical guide to S...

Full description

Saved in:
Bibliographic Details
Main Authors: Ana Victoria Ponce‐Bobadilla, Vanessa Schmitt, Corinna S. Maier, Sven Mensing, Sven Stodtmann
Format: Article
Language:English
Published: Wiley 2024-11-01
Series:Clinical and Translational Science
Online Access:https://doi.org/10.1111/cts.70056
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850144726647308288
author Ana Victoria Ponce‐Bobadilla
Vanessa Schmitt
Corinna S. Maier
Sven Mensing
Sven Stodtmann
author_facet Ana Victoria Ponce‐Bobadilla
Vanessa Schmitt
Corinna S. Maier
Sven Mensing
Sven Stodtmann
author_sort Ana Victoria Ponce‐Bobadilla
collection DOAJ
description Abstract Despite increasing interest in using Artificial Intelligence (AI) and Machine Learning (ML) models for drug development, effectively interpreting their predictions remains a challenge, which limits their impact on clinical decisions. We address this issue by providing a practical guide to SHapley Additive exPlanations (SHAP), a popular feature‐based interpretability method, which can be seamlessly integrated into supervised ML models to gain a deeper understanding of their predictions, thereby enhancing their transparency and trustworthiness. This tutorial focuses on the application of SHAP analysis to standard ML black‐box models for regression and classification problems. We provide an overview of various visualization plots and their interpretation, available software for implementing SHAP, and highlight best practices, as well as special considerations, when dealing with binary endpoints and time‐series models. To enhance the reader's understanding for the method, we also apply it to inherently explainable regression models. Finally, we discuss the limitations and ongoing advancements aimed at tackling the current drawbacks of the method.
format Article
id doaj-art-a531fbaf62f44e0aabb7b301a4749c92
institution OA Journals
issn 1752-8054
1752-8062
language English
publishDate 2024-11-01
publisher Wiley
record_format Article
series Clinical and Translational Science
spelling doaj-art-a531fbaf62f44e0aabb7b301a4749c922025-08-20T02:28:16ZengWileyClinical and Translational Science1752-80541752-80622024-11-011711n/an/a10.1111/cts.70056Practical guide to SHAP analysis: Explaining supervised machine learning model predictions in drug developmentAna Victoria Ponce‐Bobadilla0Vanessa Schmitt1Corinna S. Maier2Sven Mensing3Sven Stodtmann4AbbVie Deutschland GmbH & Co. KG Ludwigshafen GermanyAbbVie Deutschland GmbH & Co. KG Ludwigshafen GermanyAbbVie Deutschland GmbH & Co. KG Ludwigshafen GermanyAbbVie Deutschland GmbH & Co. KG Ludwigshafen GermanyAbbVie Deutschland GmbH & Co. KG Ludwigshafen GermanyAbstract Despite increasing interest in using Artificial Intelligence (AI) and Machine Learning (ML) models for drug development, effectively interpreting their predictions remains a challenge, which limits their impact on clinical decisions. We address this issue by providing a practical guide to SHapley Additive exPlanations (SHAP), a popular feature‐based interpretability method, which can be seamlessly integrated into supervised ML models to gain a deeper understanding of their predictions, thereby enhancing their transparency and trustworthiness. This tutorial focuses on the application of SHAP analysis to standard ML black‐box models for regression and classification problems. We provide an overview of various visualization plots and their interpretation, available software for implementing SHAP, and highlight best practices, as well as special considerations, when dealing with binary endpoints and time‐series models. To enhance the reader's understanding for the method, we also apply it to inherently explainable regression models. Finally, we discuss the limitations and ongoing advancements aimed at tackling the current drawbacks of the method.https://doi.org/10.1111/cts.70056
spellingShingle Ana Victoria Ponce‐Bobadilla
Vanessa Schmitt
Corinna S. Maier
Sven Mensing
Sven Stodtmann
Practical guide to SHAP analysis: Explaining supervised machine learning model predictions in drug development
Clinical and Translational Science
title Practical guide to SHAP analysis: Explaining supervised machine learning model predictions in drug development
title_full Practical guide to SHAP analysis: Explaining supervised machine learning model predictions in drug development
title_fullStr Practical guide to SHAP analysis: Explaining supervised machine learning model predictions in drug development
title_full_unstemmed Practical guide to SHAP analysis: Explaining supervised machine learning model predictions in drug development
title_short Practical guide to SHAP analysis: Explaining supervised machine learning model predictions in drug development
title_sort practical guide to shap analysis explaining supervised machine learning model predictions in drug development
url https://doi.org/10.1111/cts.70056
work_keys_str_mv AT anavictoriaponcebobadilla practicalguidetoshapanalysisexplainingsupervisedmachinelearningmodelpredictionsindrugdevelopment
AT vanessaschmitt practicalguidetoshapanalysisexplainingsupervisedmachinelearningmodelpredictionsindrugdevelopment
AT corinnasmaier practicalguidetoshapanalysisexplainingsupervisedmachinelearningmodelpredictionsindrugdevelopment
AT svenmensing practicalguidetoshapanalysisexplainingsupervisedmachinelearningmodelpredictionsindrugdevelopment
AT svenstodtmann practicalguidetoshapanalysisexplainingsupervisedmachinelearningmodelpredictionsindrugdevelopment