Fully distributed PageRank computation with exponential convergence
arXiv:1705.099278 citationsh-index: 21
Originality Incremental advance
AI Analysis
It addresses the need for efficient, low-storage distributed PageRank computation for large-scale networks.
The paper proposes a fully distributed PageRank algorithm that converges in expectation with exponential rate and requires only two scalar values per page, verified through experiments.
This work studies a fully distributed algorithm for computing the PageRank vector, which is inspired by the Matching Pursuit and features: 1) a fully distributed implementation 2) convergence in expectation with exponential rate 3) low storage requirement (two scalar values per page). Illustrative experiments are conducted to verify the findings.