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. 1304.8132
61
11
v1v2 (latest)

Local Graph Clustering Beyond Cheeger's Inequality

30 April 2013
Zeyuan Allen-Zhu
Silvio Lattanzi
Vahab Mirrokni
ArXiv (abs)PDFHTML
Abstract

Motivated by applications of large-scale graph clustering, we study random-walk-based LOCAL algorithms whose running times depend only on the size of the output cluster, rather than the entire graph. All previously known such algorithms guarantee an output conductance of O~(ϕ(A))\tilde{O}(\sqrt{\phi(A)})O~(ϕ(A)​) when the target set AAA has conductance ϕ(A)∈[0,1]\phi(A)\in[0,1]ϕ(A)∈[0,1]. In this paper, we improve it to \tilde{O}\bigg( \min\Big\{\sqrt{\phi(A)}, \frac{\phi(A)}{\sqrt{\mathsf{Conn}(A)}} \Big\} \bigg)\enspace, where the internal connectivity parameter Conn(A)∈[0,1]\mathsf{Conn}(A) \in [0,1]Conn(A)∈[0,1] is defined as the reciprocal of the mixing time of the random walk over the induced subgraph on AAA. For instance, using Conn(A)=Ω(λ(A)/log⁡n)\mathsf{Conn}(A) = \Omega(\lambda(A) / \log n)Conn(A)=Ω(λ(A)/logn) where λ\lambdaλ is the second eigenvalue of the Laplacian of the induced subgraph on AAA, our conductance guarantee can be as good as O~(ϕ(A)/λ(A))\tilde{O}(\phi(A)/\sqrt{\lambda(A)})O~(ϕ(A)/λ(A)​). This builds an interesting connection to the recent advance of the so-called improved Cheeger's Inequality [KKL+13], which says that global spectral algorithms can provide a conductance guarantee of O(ϕopt/λ3)O(\phi_{\mathsf{opt}}/\sqrt{\lambda_3})O(ϕopt​/λ3​​) instead of O(ϕopt)O(\sqrt{\phi_{\mathsf{opt}}})O(ϕopt​​). In addition, we provide theoretical guarantee on the clustering accuracy (in terms of precision and recall) of the output set. We also prove that our analysis is tight, and perform empirical evaluation to support our theory on both synthetic and real data. It is worth noting that, our analysis outperforms prior work when the cluster is well-connected. In fact, the better it is well-connected inside, the more significant improvement (both in terms of conductance and accuracy) we can obtain. Our results shed light on why in practice some random-walk-based algorithms perform better than its previous theory, and help guide future research about local clustering.

View on arXiv
Comments on this paper