Enhancing Temporal Link Prediction with HierTKG: A Hierarchical Temporal Knowledge Graph Framework

TL;DR

HierTKG is a hierarchical temporal knowledge graph framework that combines temporal graph networks and pooling techniques to improve rumor propagation prediction on social media.

Contribution

It introduces a novel hierarchical modeling approach that captures multi-scale rumor dynamics, enhancing temporal link prediction accuracy in misinformation spread.

Findings

Achieves high MRR scores on ICEWS14 and WikiData datasets.

Demonstrates robustness on noisy datasets like PHEME.

Provides scalable real-time rumor prediction capabilities.

Abstract

The rapid spread of misinformation on social media, especially during crises, challenges public decision-making. To address this, we propose HierTKG, a framework combining Temporal Graph Networks (TGN) and hierarchical pooling (DiffPool) to model rumor dynamics across temporal and structural scales. HierTKG captures key propagation phases, enabling improved temporal link prediction and actionable insights for misinformation control. Experiments demonstrate its effectiveness, achieving an MRR of 0.9845 on ICEWS14 and 0.9312 on WikiData, with competitive performance on noisy datasets like PHEME (MRR: 0.8802). By modeling structured event sequences and dynamic social interactions, HierTKG adapts to diverse propagation patterns, offering a scalable and robust solution for real-time analysis and prediction of rumor spread, aiding proactive intervention strategies.