Evaluating XAI techniques under class imbalance using CPRD data
These findings highlight a critical vulnerability in current XAI techniques, i.e., their interpretability are significantly affected under skewed class distributions, which is common in medical data and emphasises the importance of consistent model explanations for trustworthy ML deployment in healthcare.
Authors
- David J. Brown (ORCID: https://orcid.org/0000-0002-1677-7485)
- Emma O’Dowd (ORCID: https://orcid.org/0000-0001-6904-1327)
- Jaspreet Kaur (ORCID: https://orcid.org/0000-0002-0603-037X)
- Teena Rai (ORCID: https://orcid.org/0000-0003-2299-0271)
- David Baldwin (ORCID: https://orcid.org/0000-0001-8410-7160)
- Jun He (ORCID: https://orcid.org/0000-0002-5616-4691)
- Yuan Shen (ORCID: https://orcid.org/0000-0002-2697-4239)
Institutions
- King Fahd University of Petroleum and Minerals (SA)
- University of Nottingham (GB)
- Nottingham Trent University (GB)
Publication Details
- Journal
- Frontiers in Artificial Intelligence
- Published
- 2025-11-13
- DOI
- https://doi.org/10.3389/frai.2025.1682919
- Primary Topic
- Imbalanced Data Classification Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00