Fast approximation of centrality and distances in hyperbolic graphs

TL;DR

This paper introduces linear-time algorithms for approximating vertex eccentricities and distances in hyperbolic graphs with small additive errors, leveraging properties of hyperbolic geometry to efficiently analyze large real-world networks.

Contribution

The paper presents novel linear-time approximation algorithms for eccentricities and distance matrices in hyperbolic graphs, exploiting their geometric properties for efficient computation.

Findings

Algorithms achieve linear time with small additive error bounds.

Constructed shortest path trees approximate eccentricities within cδ.

Empirical results on real-world networks outperform theoretical bounds.

Abstract

We show that the eccentricities (and thus the centrality indices) of all vertices of a $\delta$-hyperbolic graph $G=(V,E)$ can be computed in linear time with an additive one-sided error of at most $c\delta$, i.e., after a linear time preprocessing, for every vertex $v$ of $G$ one can compute in $O(1)$ time an estimate $\hat{e}(v)$ of its eccentricity $ecc_G(v)$ such that $ecc_G(v)\leq \hat{e}(v)\leq ecc_G(v)+ c\delta$ for a small constant $c$. We prove that every $\delta$-hyperbolic graph $G$ has a shortest path tree, constructible in linear time, such that for every vertex $v$ of $G$, $ecc_G(v)\leq ecc_T(v)\leq ecc_G(v)+ c\delta$. These results are based on an interesting monotonicity property of the eccentricity function of hyperbolic graphs: the closer a vertex is to the center of $G$, the smaller its eccentricity is. We also show that the distance matrix of $G$ with an additive one-sided error of at most $c'\delta$ can be computed in $O(|V|^2\log^2|V|)$ time, where $c'< c$ is a small constant. Recent empirical studies show that many real-world graphs (including Internet application networks, web networks, collaboration networks, social networks, biological networks, and others) have small hyperbolicity. So, we analyze the performance of our algorithms for approximating centrality and distance matrix on a number of real-world networks. Our experimental results show that the obtained estimates are even better than the theoretical bounds.