Pattern Recognition· 2026Q1
Multi-View tensor factorization for dynamic attributed graph embedding with graph-aware features
- 0citations
- Q1SCImago
- 2026year
Short summary
A new Multi-View Tensor Factorization (MV-TF) method effectively integrates topology, attributes, and graph-aware features to model dynamic attributed graphs, outperforming state-of-the-art in node classification.
AI-generated from the title and abstract; the full text is not read.
Key points
- Introduces Multi-View Tensor Factorization (MV-TF) for dynamic attributed graph embedding.
- Jointly models topology, attributes, and graph-aware contextual features in a shared latent space.
- Explicitly captures coupling between different data views and uses a sparse subspace constraint.
- Demonstrates superior performance over state-of-the-art methods in node classification tasks.
AI-generated from the title and abstract; the full text is not read.
Abstract
Learning node representations in dynamic graphs has garnered significant attention in recent years. However, existing methods face challenges in effectively integrating topology and node attributes within individual snapshots while simultaneously capturing the evolution of graph structure and attributes across snapshots. Motivated by the above challenges, we formulate unsupervised dynamic attributed graph embedding as a coupled multi-view tensor learning problem, where topology, attributes, and graph-aware contextual features are jointly modeled by Multi-View Tensor Factorization (MV-TF) within a shared latent space. The framework explicitly models the coupling between views and incorporates a sparse subspace constraint to enhance embedding quality. Extensive experimental results on node classification tasks demonstrate that our method outperforms state-of-the-art approaches. These results highlight its ability to effectively mine the evolution patterns of dynamic graphs and provide valuable insights for related analyses.
The authors' abstract, as published at the source. Pattern Recognition, 2026 · DOI ↗
Continue with a free account
Ask the paper: 3 free questions a day about this paper; save it, get its citation, new summaries every day for your field. Takeaways are Premium.
Continue free on the webSign in with Google or Apple; no card needed. You come back to this paper.
On your phone:
Field: Computational Mathematics
Computational MathematicsMathematics