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

• Published in: Clinical and translational science Vol. 17; no. 11; pp. e70056 - n/a
• Main Authors: Ponce‐Bobadilla, Ana Victoria, Schmitt, Vanessa, Maier, Corinna S., Mensing, Sven, Stodtmann, Sven
• Format: Journal Article
• Language: English
• Published: United States John Wiley & Sons, Inc 01.11.2024; John Wiley and Sons Inc; Wiley
• Subjects: Artificial intelligence; Artificial Intelligence and Machine Learning; Collaboration; Combination therapy; Drug development; Drug Development - methods; Game theory; Humans; Learning algorithms; Machine learning; Methods; Neural networks; Predictions; Regression analysis; Response rates; Software; Supervised Machine Learning; Tutorial
• ISSN: 1752-8054; 1752-8062; 1752-8062
• DOI: 10.1111/cts.70056

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.