.png){kind=icon}
.png){kind=icon}
Return
HyperDVM: A hypergraph model with dual-view selection mechanism
.png){kind=icon}
PRE
AI
DOI：10.1016/j.ins.2025.122980.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Hypergraphs provide a powerful framework for representing higher-order interactions in complex systems across bioinformatics, social networks, communication systems, and knowledge graphs. Yet existing hypergraph evolution models typically depend on node attributes or fixed parameters, limiting their ability to capture real-world dynamics driven by resource constraints and the self-organization of higher-order structures. We propose HyperDVM, a dual-perspective hyper-graph evolution model that operates on weight-adjusted probabilities to jointly select nodes and hyperedges. By integrating a (k, q)-core decomposition to identify highly connected cores, HyperDVM systematically simulates both the incremental evolution of hypergraphs and the emergence of dense subgraphs. A parameterized weight-adjustment function controls node hyperdegree, hyperedge size, and their coupling, enabling the model to flexibly reproduce a range of higher-order interaction patterns under finite-node and finite-hyperedge constraints. Validation on empirical hypergraph datasets shows strong structural reconstruction performance and the capacity to emulate diverse evolutionary mechanisms. Complementary analyses on synthetic hypergraphs illuminate how model parameters affect network scale, heterogeneity, and key structural metrics. Together, these results establish HyperDVM as a versatile theoretical framework for understanding, designing, and controlling the generation and evolution of higher-order networks.
Keywords:
Hypergraph
Complex Systems
Dynamic Evolution
Joint Selection Mechanism
Journal
IF:
6.8
Papers:
540
Citations:
6.2W