Correlation clustering with stochastic labellings

Nicola Rebagliati, S. Rota Bulo, M. Pelillo

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

2 Citations (Scopus)

Abstract

Correlation clustering is the problem of finding a crisp partition of the vertices of a correlation graph in such a way as to minimize the disagreements in the cluster assignments. In this paper, we discuss a relaxation to the original problem setting which allows probabilistic assignments of vertices to labels. By so doing, overlapping clusters can be captured. We also show that a known optimization heuristic can be applied to the problem formulation, but with the automatic selection of the number of classes. Additionally, we propose a simple way of building an ensemble of agreement functions sampled from a reproducing kernel Hilbert space, which allows to apply correlation clustering without the empirical estimation of pairwise correlation values.
Original languageEnglish
Title of host publicationSimilarity-Based Pattern Recognition
Subtitle of host publicationSIMBAD 2013
Place of PublicationBerlin, Heidelberg
PublisherSpringer
Pages120-133
ISBN (Electronic)978-3-642-39140-8
ISBN (Print)978-3-642-39139-2
DOIs
Publication statusPublished - 2013
MoE publication typeNot Eligible
Event2nd International Workshop on Similarity-Based Pattern Analysis and Recognition, SIMBAD 2013 - York, United Kingdom
Duration: 3 Jul 20135 Jul 2013

Publication series

SeriesLecture Notes in Computer Science
Volume7953
ISSN0302-9743

Conference

Conference2nd International Workshop on Similarity-Based Pattern Analysis and Recognition, SIMBAD 2013
Abbreviated titleSIMBAD 2013
CountryUnited Kingdom
CityYork
Period3/07/135/07/13

Keywords

  • Baum-Eagon inequality
  • correlation clustering
  • ensemble clustering
  • stochastic labelling

Fingerprint Dive into the research topics of 'Correlation clustering with stochastic labellings'. Together they form a unique fingerprint.

Cite this