• Media type: Text; E-Article; Electronic Conference Proceeding
  • Title: Cluster Editing with Overlapping Communities
  • Contributor: Arrighi, Emmanuel [Author]; Bentert, Matthias [Author]; Drange, Pål Grønås [Author]; Sullivan, Blair D. [Author]; Wolf, Petra [Author]
  • Published: Schloss Dagstuhl – Leibniz-Zentrum für Informatik, 2023
  • Language: English
  • DOI: https://doi.org/10.4230/LIPIcs.IPEC.2023.2
  • Keywords: graph modification ; correlation clustering ; vertex splitting ; NP-hardness ; parameterized algorithm
  • Origination:
  • Footnote: Diese Datenquelle enthält auch Bestandsnachweise, die nicht zu einem Volltext führen.
  • Description: Cluster Editing, also known as correlation clustering, is a well-studied graph modification problem. In this problem, one is given a graph and allowed to perform up to k edge additions and deletions to transform it into a cluster graph, i.e., a graph consisting of a disjoint union of cliques. However, in real-world networks, clusters are often overlapping. For example, in social networks, a person might belong to several communities - e.g. those corresponding to work, school, or neighborhood. Another strong motivation comes from language networks where trying to cluster words with similar usage can be confounded by homonyms, that is, words with multiple meanings like "bat". The recently introduced operation of vertex splitting is one natural approach to incorporating such overlap into Cluster Editing. First used in the context of graph drawing, this operation allows a vertex v to be replaced by two vertices whose combined neighborhood is the neighborhood of v (and thus v can belong to more than one cluster). The problem of transforming a graph into a cluster graph using at most k edge additions, edge deletions, or vertex splits is called Cluster Editing with Vertex Splitting and is known to admit a polynomial kernel with respect to k and an O(9^{k²} + n + m)-time (parameterized) algorithm. However, it was not known whether the problem is NP-hard, a question which was originally asked by Abu-Khzam et al. [Combinatorial Optimization, 2018]. We answer this in the affirmative. We further give an improved algorithm running in O(2^{7klog k} + n + m) time.
  • Access State: Open Access