Distributed and incremental clustering based on weighted affinity propagation

Xiangliang Zhang*, Cyril Furtlehner, Michèle Sebag

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

5 Scopus citations

Abstract

A new clustering algorithm Affinity Propagation (AP) is hindered by its quadratic complexity. The Weighted Affinity Propagation (WAP) proposed in this paper is used to eliminate this limitation, support two scalable algorithms. Distributed AP clustering handles large datasets by merging the exemplars learned from subsets. Incremental AP extends AP to online clustering of data streams. The paper validates all proposed algorithms on benchmark and on real-world datasets. Experimental results show that the proposed approaches offer a good trade-off between computational effort and performance.

Original languageEnglish (US)
Title of host publicationSTAIRS 2008. Proceedings of the Fourth Starting AI Researchers' Symposium
PublisherIOS Press
Pages199-210
Number of pages12
Edition1
ISBN (Print)9781586038939
DOIs
StatePublished - Jan 1 2008

Publication series

NameFrontiers in Artificial Intelligence and Applications
Number1
Volume179
ISSN (Print)0922-6389

Keywords

  • Affinity Propagation
  • Data Clustering
  • Data Streaming
  • K-centers

ASJC Scopus subject areas

  • Artificial Intelligence

Fingerprint

Dive into the research topics of 'Distributed and incremental clustering based on weighted affinity propagation'. Together they form a unique fingerprint.

Cite this