Skip to content

Hashtag networks

Hashtags are metadata tags used on social media platforms to categorize and organize content. In the context of misinformation detection, hashtag networks capture direct and indirect relationships among hashtags based on their co-occurrence in posts about the same news article or claim.

Key insights

  • Hashtag credibility signals: Fake news tends to cluster with hashtags used primarily for spreading misinformation, while true news uses hashtags associated with reliable content
  • Co-occurrence as graph structure: Hashtags appearing in the same posts form edges in a network; paths through this network reveal indirect relationships that may not be visible from direct co-occurrence alone
  • Language-independent signal: Hashtag analysis requires only counting co-occurrence patterns, making it applicable across languages without NLP preprocessing
  • Early detection advantage: Hashtag credibility can be assessed with limited propagation data, enabling rapid classification before news reaches viral scale

Complementary to other approaches

Hashtag-based detection works alongside: - Propagation-based detection (leverages dissemination structure but often requires longer cascade histories) - Graph-based detection (both use graphs but hashtag graphs are simpler and faster to construct than user-interaction graphs) - Content-based detection (hashtags provide signals orthogonal to linguistic analysis)

Key papers in this wiki