Learning Heuristics to Reduce the Overestimation of Bipartite Graph Edit Distance Approximation

dc.accessRightsAnonymous
dc.audienceScience
dc.contributor.authorFerrer, Miguel
dc.contributor.authorSerratosa, Francesco
dc.contributor.authorRiesen, Kaspar
dc.contributor.editorPerner, Petra
dc.date.accessioned2015-10-08T10:17:57Z
dc.date.available2015-10-08T10:17:57Z
dc.date.issued2015
dc.description.abstractIn data mining systems, which operate on complex data with structural relationships, graphs are often used to represent the basic objects under study. Yet, the high representational power of graphs is also accompanied by an increased complexity of the associated algorithms. Exact graph similarity or distance, for instance, can be computed in exponential time only. Recently, an algorithmic framework that allows graph dissimilarity computation in cubic time with respect to the number of nodes has been presented. This fast computation is at the expense, however, of generally overestimating the true distance. The present paper introduces six different post-processing algorithms that can be integrated in this suboptimal graph distance framework. These novel extensions aim at improving the overall distance quality while keeping the low computation time of the approximation. An experimental evaluation clearly shows that the proposed heuristics substantially reduce the overestimation in the existing approximation framework while the computation time remains remarkably low.
dc.event.end2015-07-21
dc.event.start2015-07-20
dc.identifier.isbn978-3-319-21023-0
dc.identifier.urihttp://hdl.handle.net/11654/10151
dc.language.isoen
dc.publisherSpringer
dc.relation.ispartofMachine Learning and Data Mining in Pattern Recognition - 11th International Conference, MLDM 2015, Hamburg, Germany, July 20-21, 2015
dc.relation.ispartofseriesLecture Notes in Computer Science
dc.spatialHamburg
dc.titleLearning Heuristics to Reduce the Overestimation of Bipartite Graph Edit Distance Approximation
dc.type04 - Beitrag Sammelband oder Konferenzschrift
dspace.entity.typePublication
fhnw.InventedHereYes
fhnw.IsStudentsWorkno
fhnw.PublishedSwitzerlandNo
fhnw.ReviewTypeAnonymous ex ante peer review of a complete publication
fhnw.affiliation.hochschuleHochschule für Wirtschaftde_CH
fhnw.affiliation.institutInstitut für Wirtschaftsinformatikde_CH
fhnw.pagination17-31
fhnw.publicationStatePublished
fhnw.seriesNumber9166
relation.isAuthorOfPublicationd761e073-1612-4d22-8521-65c01c19f97a
relation.isAuthorOfPublication.latestForDiscoveryd761e073-1612-4d22-8521-65c01c19f97a
Dateien
Lizenzbündel
Gerade angezeigt 1 - 1 von 1
Lade...
Vorschaubild
Name:
license.txt
Größe:
2.94 KB
Format:
Item-specific license agreed upon to submission
Beschreibung: