PHEME Misinformation Graphs: Twitter Conversation Cascades for GNNs

Graph representations of Twitter conversation cascades from the PHEME 9-Event rumour dataset. The dataset is designed for Graph-Level Binary Classification tasks to detect misinformation spreading dynamics using Graph Neural Networks. Each graph represents a single breaking-news conversation thread, with nodes as tweets and edges as reply-to relationships.

Use Cases

• Training Graph Neural Networks for misinformation detection based on reply-to network structures.
• Analyzing the spread dynamics of rumours in breaking-news events based on conversation cascades.
• Benchmarking graph-level binary classification models for social media content.
• Studying the structural patterns of misinformation propagation on Twitter.

Strengths

• Dataset is specifically designed for Graph-Level Binary Classification tasks.
• Graphs represent real Twitter conversation cascades from a known rumour dataset (PHEME 9-Event).
• Nodes and edges are clearly defined as tweets and reply-to relationships.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and sample data are unknown, which may limit suitability assessment.
• License information is unknown, which could restrict usage.

Provenance

Source

PHEME 9-Event rumour dataset.

Collection Method

Mathematical graph representations of Twitter conversation cascades.

Freshness

Last updated 2026-05-04 23:28:24; freshness should be verified.