Link prediction with hyperbolic geometry

2019/03/31 by Maksim Kitsak, Ivan Voitalov, Dmitri Krioukov
Computer Science · Mathematics · Physics and Astronomy · #Advanced Graph Neural Networks #Algorithm #Artificial intelligence #Combinatorics #Complex Network Analysis Techniques #Computer science #Differential geometry #Embedding #Geometry #Hyperbolic coordinates #Hyperbolic equilibrium point #Hyperbolic function #Hyperbolic geometry #Hyperbolic manifold #Hyperbolic space #Hyperbolic triangle #Inference #Link (geometry) #Mathematics #Opinion Dynamics and Social Influence #cs.SI #physics.soc-ph

paper · pdf · doi:10.1103/physrevresearch.2.043113

published as Phys. Rev. Research 2, 043113 (2020)

openalex publication_date 2020/10/21 · arxiv created 2020/11/02 · arxiv updated 2020/11/03 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/06

Abstract

The paper finds that the accuracy of link prediction in networks using latent hyperbolic geometry depends on how accurately coordinates of nodes in this geometry are inferred, and develops a hyperbolic coordinate inference method for network embedding.

Citations

• Generalized Communities in Networks
• Hyperbolic geometry of complex networks
• Missing and spurious interactions and the reconstruction of complex networks
• Birds of a Feather: Homophily in Social Networks
• Friends and neighbors on the Web
• Multiscale unfolding of real networks by geometric renormalization
• A nonuniform popularity-similarity optimization (nPSO) model to efficiently generate realistic complex networks with communities
• Scale-free networks well done
• Machine learning meets complex networks via coalescent embedding in the hyperbolic space