@inproceedings{patularu-2026-bionlp,
    title = "{B}io{NLP} at {\#}{SMM}4{H}-{H}ea{RD} 2026 Task 3 Estimating Flu Vaccine Effectiveness: A Temporal-Aware Fine-Tuning and Similarity-Based Few-Shot Prompting Approach",
    author = "Patularu, Irina",
    editor = "Lopez-Garcia, Guillermo  and
      Gonzalez-Hernandez, Graciela",
    booktitle = "Proceedings of the 11th Social Media Mining for Health Research and Applications ({SMM}4{H}-{H}ea{RD} 2026) Workshop and Shared Tasks",
    month = jul,
    year = "2026",
    address = "San Diego, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl-workshops/2026.smm4h-1.41/",
    pages = "252--259",
    ISBN = "979-8-89176-432-3",
    abstract = "This paper presents our systems for the SMM4H 2026 shared task on flu-related tweetclassification across two subtasks: flu vaccination status and flu test outcome classification. For each subtask, we evaluate two approaches: fine-tuning BERTweet-large with atemporal-aware architecture, cross-validation ensembling, and regularization techniques, anda GPT-4o few-shot prompting system with similarity-based dynamic example retrieval,chain-of-thought reasoning and contrastive label ranking. Fine-tuning proves superior for theflu vaccination subtask (micro-F1: 87.90{\%}), where sufficient and relatively balanced training datais available, while few-shot prompting performs better for the flu test subtask (micro-F1: 95.74{\%}), where limited and heavily imbalanced training data renders fine-tuning less effective."
}