Statistical reliability and path diversity based PageRank algorithm improvements

TL;DR

This paper proposes enhancements to the PageRank algorithm by incorporating statistical reliability and path diversity, aiming to improve ranking accuracy through dynamic adjustments based on local graph properties.

Contribution

It introduces novel concepts of statistical reliability and path diversity into PageRank, enabling dynamic modifications of node scores and damping factors.

Findings

Improved ranking accuracy demonstrated through examples.

Dynamic adjustments enhance PageRank's adaptability.

Simulation results support the effectiveness of proposed ideas.

Abstract

In this paper we present new improvement ideas of the original PageRank algorithm. The first idea is to introduce an evaluation of the statistical reliability of the ranking score of each node based on the local graph property and the second one is to introduce the notion of the path diversity. The path diversity can be exploited to dynamically modify the increment value of each node in the random surfer model or to dynamically adapt the damping factor. We illustrate the impact of such modifications through examples and simple simulations.