oai:arXiv.org:2412.18779
Computer Science
2024
01/01/2025
Long COVID continues to challenge public health by affecting a significant segment of individuals who have recovered from acute SARS-CoV-2 infection yet endure prolonged and often debilitating symptoms.
Social media has emerged as a vital resource for those seeking real-time information, peer support, and validating their health concerns related to Long COVID.
This paper examines recent works focusing on mining, analyzing, and interpreting user-generated content on social media platforms such as Twitter, Reddit, Facebook, and YouTube to capture the broader discourse on persistent post-COVID conditions.
A novel transformer-based zero-shot learning approach serves as the foundation for classifying research papers in this area into four primary categories: Clinical or Symptom Characterization, Advanced NLP or Computational Methods, Policy, Advocacy, or Public Health Communication, and Online Communities and Social Support.
This methodology showcases the adaptability of advanced language models in categorizing research papers without predefined training labels, thus enabling a more rapid and scalable assessment of existing literature.
This review highlights the multifaceted nature of Long COVID research, where computational techniques applied to social media data reveal insights into narratives of individuals suffering from Long COVID.
This review also demonstrates the capacity of social media analytics to inform clinical practice and contribute to policy-making related to Long COVID.
Thakur, Nirmalya, 2024, Integrating Zero-Shot Classification to Advance Long COVID Literature: A Systematic Social Media-Centered Review