Abstract:We study the implications of the modeling choice to use a graph, instead of a hypergraph, to represent real-world interconnected systems whose constituent relationships are of higher order by nature. Such a modeling choice typically involves an underlying projection process that maps the original hypergraph onto a graph, and is common in graph-based analysis. While hypergraph projection can potentially lead to loss of higher-order relations, there exists very limited studies on the consequences of doing so, as well as its remediation. This work fills this gap by doing two things: (1) we develop analysis based on graph and set theory, showing two ubiquitous patterns of hyperedges that are root to structural information loss in all hypergraph projections; we also quantify the combinatorial impossibility of recovering the lost higher-order structures if no extra help is provided; (2) we still seek to recover the lost higher-order structures in hypergraph projection, and in light of (1)'s findings we propose to relax the problem into a learning-based setting. Under this setting, we develop a learning-based hypergraph reconstruction method based on an important statistic of hyperedge distributions that we find. Our reconstruction method is evaluated on 8 real-world datasets under different settings, and exhibits consistently good performance. We also demonstrate benefits of the reconstructed hypergraphs via use cases of protein rankings and link predictions.

Hypernetwork Science via High-Order Hypergraph Walks

Understanding High-Order Network Structure using Permissible Walks on Attributed Hypergraphs

Flow-Based Community Detection in Hypergraphs

Hypernetwork Science: From Multidimensional Networks to Computational Topology

From unbiased to maximal entropy random walks on hypergraphs

From unbiased to maximal-entropy random walks on hypergraphs

A framework to generate hypergraphs with community structure

Framework to generate hypergraphs with community structure

Formal Concept Lattice Representations and Algorithms for Hypergraphs

Hyperlink communities in higher-order networks

CAT-Walk: Inductive Hypergraph Learning via Set Walks

Weighted projected networks: mapping hypergraphs to networks

Identifying vital nodes through augmented random walks on higher-order networks

From Graphs to Hypergraphs: Hypergraph Projection and its Remediation

Uncovering the hidden core-periphery structure in hyperbolic networks

Structure and inference in hypergraphs with node attributes

Two Layer Walk: A Community-Aware Graph Embedding

Community Detection in Large Hypergraphs

Hypergraph Partitioning for Community Discovery in Complex Network