Finding density-based subspace clusters in graphs with feature vectors

Stephan Gunnemann, Brigitte Boden, Thomas Seidl

Research output: Contribution to conferencePaperpeer-review

Abstract

Data sources representing attribute information in combination with network information are widely available in today's applications. To realize the full potential for knowledge extraction, mining techniques like clustering should consider both information types simultaneously. Recent clustering approaches combine subspace clustering with dense subgraph mining to identify groups of objects that are similar in subsets of their attributes as well as densely connected within the network. While those approaches successfully circumvent the problem of full-space clustering, their limited cluster definitions are restricted to clusters of certain shapes. In this work, we introduce a density-based cluster definition taking the attribute similarity in subspaces and the graph density into account. This novel cluster model enables us to detect clusters of arbitrary shape and size. We avoid redundancy in the result by selecting only the most interesting non-redundant clusters. Based on this model, we introduce the clustering algorithm DB-CSC. In the experimental evaluation we demonstrate the strength of DB-CSC in comparison to related clustering approaches.

Original languageEnglish
Pages20-27
Number of pages8
StatePublished - 2011
Externally publishedYes
EventSymposium on Learning, Knowledge, and Adaptivity 2011, LWA 2011 - Magdeburg, Germany
Duration: 28 Sep 201130 Sep 2011

Conference

ConferenceSymposium on Learning, Knowledge, and Adaptivity 2011, LWA 2011
Country/TerritoryGermany
CityMagdeburg
Period28/09/1130/09/11

Fingerprint

Dive into the research topics of 'Finding density-based subspace clusters in graphs with feature vectors'. Together they form a unique fingerprint.

Cite this