A Hybrid TGN-SEAL Model for Dynamic Graph Link Prediction

TL;DR

The paper addresses dynamic link prediction in continuous-time, sparse graphs by blending Temporal Graph Networks (TGN) with SEAL-style enclosing subgraphs. It introduces TGN-SEAL, which uses $k$-hop enclosing subgraphs around candidate links (notably $k=2$ and $k=3$) and a subgraph-level GNN on top of temporal embeddings to predict links. On the Reality Mining call-detail dataset, TGN-SEAL achieves higher mean average precision than strong baselines, with approximately $2.6\%$ gains on unseen nodes and $1.6\%$ gains on seen nodes, illustrating the value of combining local topology with temporal dynamics. The approach reveals a practical trade-off between predictive performance and computational overhead, and points to future work on scalability and extension to heterogeneous networks.

Abstract

Predicting links in sparse, continuously evolving networks is a central challenge in network science. Conventional heuristic methods and deep learning models, including Graph Neural Networks (GNNs), are typically designed for static graphs and thus struggle to capture temporal dependencies. Snapshot-based techniques partially address this issue but often encounter data sparsity and class imbalance, particularly in networks with transient interactions such as telecommunication call detail records (CDRs). Temporal Graph Networks (TGNs) model dynamic graphs by updating node embeddings over time; however, their predictive accuracy under sparse conditions remains limited. In this study, we improve the TGN framework by extracting enclosing subgraphs around candidate links, enabling the model to jointly learn structural and temporal information. Experiments on a sparse CDR dataset show that our approach increases average precision by 2.6% over standard TGNs, demonstrating the advantages of integrating local topology for robust link prediction in dynamic networks.

Figures (4)

• Figure 1: Classification of Link Prediction Methods
• Figure 2: Proposed architecture of the TGN-SEAL framework.
• Figure 3: Training loss of TGN-SEAL and baseline methods.
• Figure 4: Mean Average Precision of TGN-SEAL and baseline methods.