New TGN-SEAL Model Enhances Link Prediction in Sparse Dynamic Networks

Researchers have developed a new model called TGN-SEAL to improve link prediction in dynamic and sparse networks. This hybrid approach combines Temporal Graph Networks (TGNs) with the extraction of enclosing subgraphs around candidate links. Experiments on telecommunication call detail records and email datasets showed that TGN-SEAL increases average precision by at least 2% compared to standard TGNs, demonstrating its effectiveness in capturing both structural and temporal information for robust link prediction. AI

IMPACT This research offers a novel method for improving link prediction in dynamic and sparse networks, potentially benefiting fields that rely on analyzing evolving relational data.

RANK_REASON Academic paper detailing a new model and experimental results. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Graph Neural Networks
• Nafiseh Sadat Sajadi
• Temporal Graph Networks
• TGN-SEAL

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →