English

Personalized PageRank to a Target Node, Revisited

Data Structures and Algorithms 2020-06-25 v2

Abstract

Personalized PageRank (PPR) is a widely used node proximity measure in graph mining and network analysis. Given a source node ss and a target node tt, the PPR value π(s,t)\pi(s,t) represents the probability that a random walk from ss terminates at tt, and thus indicates the bidirectional importance between ss and tt. The majority of the existing work focuses on the single-source queries, which asks for the PPR value of a given source node ss and every node tVt \in V. However, the single-source query only reflects the importance of each node tt with respect to ss. In this paper, we consider the {\em single-target PPR query}, which measures the opposite direction of importance for PPR. Given a target node tt, the single-target PPR query asks for the PPR value of every node sVs\in V to a given target node tt. We propose RBS, a novel algorithm that answers approximate single-target queries with optimal computational complexity. We show that RBS improves three concrete applications: heavy hitters PPR query, single-source SimRank computation, and scalable graph neural networks. We conduct experiments to demonstrate that RBS outperforms the state-of-the-art algorithms in terms of both efficiency and precision on real-world benchmark datasets.

Cite

@article{arxiv.2006.11876,
  title  = {Personalized PageRank to a Target Node, Revisited},
  author = {Hanzhi Wang and Zhewei Wei and Junhao Gan and Sibo Wang and Zengfeng Huang},
  journal= {arXiv preprint arXiv:2006.11876},
  year   = {2020}
}

Comments

ACM SIGKDD 2020

R2 v1 2026-06-23T16:29:59.467Z