Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Heuristics for Inequality minimization in PageRank values

Subhajit Sahu

TL;DR

This work investigates inequality in PageRank distributions by measuring dispersion with the Gini coefficient $G$ and evaluating six deterministic heuristics that add edges to reduce $G$. By recomputing PageRank after each insertion, the study finds that inequality reduction is most effective on graphs with high initial $G$ (e.g., web and social networks) and that a combination of two heuristics, notably $Cxrx$ and $CxSx$, yields the strongest per-edge improvements. Gains diminish as more edges are added, and performance on low-$G$ graphs is limited, highlighting context dependence for inequality mitigation. The results offer a practical route toward more equitable PageRank distributions in large networks and motivate future exploration of randomized or alternative deterministic strategies.

Abstract

PageRank is a widely used algorithm for ranking webpages and plays a significant role in determining web traffic. This study employs the Gini coefficient, a measure of income/wealth inequality, to assess the inequality in PageRank distributions and explores six deterministic methods for reducing inequality. Our findings indicate that a combination of two distinct heuristics may present an effective strategy for minimizing inequality.

Heuristics for Inequality minimization in PageRank values

TL;DR

This work investigates inequality in PageRank distributions by measuring dispersion with the Gini coefficient and evaluating six deterministic heuristics that add edges to reduce . By recomputing PageRank after each insertion, the study finds that inequality reduction is most effective on graphs with high initial (e.g., web and social networks) and that a combination of two heuristics, notably and , yields the strongest per-edge improvements. Gains diminish as more edges are added, and performance on low- graphs is limited, highlighting context dependence for inequality mitigation. The results offer a practical route toward more equitable PageRank distributions in large networks and motivate future exploration of randomized or alternative deterministic strategies.

Abstract

PageRank is a widely used algorithm for ranking webpages and plays a significant role in determining web traffic. This study employs the Gini coefficient, a measure of income/wealth inequality, to assess the inequality in PageRank distributions and explores six deterministic methods for reducing inequality. Our findings indicate that a combination of two distinct heuristics may present an effective strategy for minimizing inequality.
Paper Structure (13 sections, 2 equations, 1 figure, 1 table, 1 algorithm)

This paper contains 13 sections, 2 equations, 1 figure, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related work
  3. Preliminaries
  4. PageRank
  5. Gini coefficient
  6. Approach
  7. Evaluation
  8. Experimental Setup
  9. System used
  10. Configuration
  11. Dataset
  12. Results
  13. Conclusion

Figures (1)

  • Figure 1: Variation of Gini coefficient (Y-axis) with edges being added (X-axis) to the graphs incrementally with six different heuristics: Cxrx, CxSx, CxSr, CRrx, CRSx, and CRSr (see Algorithm \ref{['alg:heuristic']} for the psuedocode).