Workshop Proceedings of the 19th International AAAI Conference on Web and Social Media

Workshop: #SMM4H-HeaRD 2025: Joint 10th Social Media Mining for Health and Health Real-World Data Workshop and Shared Tasks

DOI: 10.36190/2025.66

Published: 2025-06-05
RIGA at SMM4H-HeaRD 2025: Context-Enriched Classification Pipeline
Eduards Mukans, Guntis Barzdins

The following is a description of the RIGA team's submissions for the SMM4H-HeaRD 2025 Task 1: Detection of adverse drug events (ADEs) in multilingual and multi-platform social media posts. Our approach leverages Large Language Models (LLMs) and knowledge databases to design a set of informative features that enhance the fine-tuning of a sequence classification model. Experimental results demonstrate that this method improves the F1 score and leads to more balanced predictions.