Efficient Sparsification of Densely Connected Clusters

Joyentanuj Das, Suranjan De, He Sun

Keywords: densely connected components, graph sparsification, clustering

TL;DR: We present two algorithms that sparsify densely connected clusters in both undirected graphs and directed ones.

Abstract: When modelling a real-world dataset as a graph, groups of highly correlated data items correspond to densely connected vertex sets (clusters), and efficient algorithms that find these clusters have broad applications in various data analysis tasks. In this paper we study densely connected clusters in graphs and introduce two sparsification algorithms that preserve the structure of these clusters in both undirected graphs and directed ones. We show that our algorithms significantly speedup the running time of existing clustering algorithms while preserving their effectiveness.

Primary Area: learning on graphs and other geometries & topologies

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 11226