ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2306.08617
50
4
v1v2 (latest)

Multi-class Graph Clustering via Approximated Effective ppp-Resistance

14 June 2023
Shota Saito
Mark Herbster
ArXiv (abs)PDFHTML
Abstract

This paper develops an approximation to the (effective) ppp-resistance and applies it to multi-class clustering. Spectral methods based on the graph Laplacian and its generalization to the graph ppp-Laplacian have been a backbone of non-euclidean clustering techniques. The advantage of the ppp-Laplacian is that the parameter ppp induces a controllable bias on cluster structure. The drawback of ppp-Laplacian eigenvector based methods is that the third and higher eigenvectors are difficult to compute. Thus, instead, we are motivated to use the ppp-resistance induced by the ppp-Laplacian for clustering. For ppp-resistance, small ppp biases towards clusters with high internal connectivity while large ppp biases towards clusters of small "extent," that is a preference for smaller shortest-path distances between vertices in the cluster. However, the ppp-resistance is expensive to compute. We overcome this by developing an approximation to the ppp-resistance. We prove upper and lower bounds on this approximation and observe that it is exact when the graph is a tree. We also provide theoretical justification for the use of ppp-resistance for clustering. Finally, we provide experiments comparing our approximated ppp-resistance clustering to other ppp-Laplacian based methods.

View on arXiv
Comments on this paper