A Proximity Measure using Blink Model

TL;DR

This paper introduces a new graph proximity measure based on network reliability, demonstrating its superior consistency with human intuition and promising empirical results in link prediction tasks across different networks.

Contribution

It presents a novel proximity measure derived from network reliability and a deterministic approximation algorithm, outperforming existing predictors in link prediction benchmarks.

Findings

Achieves 14-35% higher accuracy than previous best methods.

More consistent with human intuition than existing measures.

Effective in coauthorship and Wikipedia network benchmarks.

Abstract

This paper proposes a new graph proximity measure. This measure is a derivative of network reliability. By analyzing its properties and comparing it against other proximity measures through graph examples, we demonstrate that it is more consistent with human intuition than competitors. A new deterministic algorithm is developed to approximate this measure with practical complexity. Empirical evaluation by two link prediction benchmarks, one in coauthorship networks and one in Wikipedia, shows promising results. For example, a single parameterization of this measure achieves accuracies that are 14-35% above the best accuracy for each graph of all predictors reported in the 2007 Liben-Nowell and Kleinberg survey.