Large-Model-Based Automatic Annotation for Health Text Mining in Social Media
International Journal of Computational and Biological Sciences · 20.04.2026 · 3 (1) · 265-276
DOI: 10.66238/ijcbs68 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Computational and Biological Sciences, online erschienen am 20.04.2026. Verfasst von Sofia Cabrera. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
The exponential growth of user-generated content on social media platforms presents an unprecedented opportunity for health text mining, enabling applications such as pharmacovigilance, syndromic surveillance, and mental health monitoring. However, the unstructured, colloquial, and noisy nature of social media text severely limits the application of traditional natural language processing techniques. Furthermore, the reliance on manual annotation to create high-quality, domain-specific training datasets constitutes a significant bottleneck, given the high costs and requirements for specialized medical knowledge. This paper comprehensively explores the application of Large Language Models to automate the annotation of health-related texts sourced from social media platforms. By formulating a sophisticated annotation framework that integrates zero-shot learning, few-shot prompting, and chain-of-thought reasoning, we investigate how large models can effectively replace or augment human annotators. The study presents a rigorous methodology for standardizing social media inputs, designing context-aware prompts, and calculating consensus scores to ensure annotation fidelity. Quantitative and qualitative analyses reveal that large models can achieve high concordance with human experts, particularly in identifying broad phenotypic expressions and adverse drug events. The findings suggest a paradigm shift in health informatics, demonstrating that automated annotation pipelines can drastically reduce dataset creation time while maintaining high standards of data quality, thereby accelerating public health research and responsive policy-making
Zitieren
APA
Cabrera, S. (2026). Large-Model-Based Automatic Annotation for Health Text Mining in Social Media. International Journal of Computational and Biological Sciences, 3(1), 265-276. https://doi.org/10.66238/ijcbs68
Vancouver
Cabrera S. Large-Model-Based Automatic Annotation for Health Text Mining in Social Media. International Journal of Computational and Biological Sciences. 2026;3(1):265-276. doi:10.66238/ijcbs68
BibTeX
@article{Cabrera2026,
author = {Cabrera, Sofia},
title = {Large-Model-Based Automatic Annotation for Health Text Mining in Social Media},
journal = {International Journal of Computational and Biological Sciences},
year = {2026},
volume = {3},
number = {1},
pages = {265--276},
doi = {10.66238/ijcbs68}
}
EUVIMED