Zero-shot learning for clinical phenotyping: Comparing LLMs and rule-based methods
Zero-shot learning with LLMs, particularly GPT-4o, offers a powerful and efficient approach for phenotyping chronic conditions from EHRs, significantly reducing the need for extensive labeled datasets while maintaining high accuracy and interpretability.
Authors
- Jorge Cerejo
- Francisca Leite (ORCID: https://orcid.org/0000-0003-2550-2616)
- Nuno Silva (ORCID: https://orcid.org/0000-0003-4216-2107)
- Bernardo Neves (ORCID: https://orcid.org/0000-0002-1559-7482)
- Mário J. Silva (ORCID: https://orcid.org/0000-0002-5452-6185)
- Simão Gonçalves
- José-Maria Moreira (ORCID: https://orcid.org/0000-0003-2420-7930)
Institutions
- Instituto de Engenharia de Sistemas e Computadores Investigação e Desenvolvimento (PT)
- University of Lisbon (PT)
- Universidade Católica Portuguesa (PT)
- Hospital da Luz (PT)
Publication Details
- Journal
- Computers in Biology and Medicine
- Published
- 2025-04-23
- DOI
- https://doi.org/10.1016/j.compbiomed.2025.110181
- Citations
- 4
- Primary Topic
- Machine Learning in Healthcare
- Type
- article
- Field-Weighted Citation Impact
- 9.92