Clustering techniques for minimizing object access time

Vlad S. I. Wietrzyk, Mehmet A. Orgun

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

Abstract

We propose three designs for clustering objects: a new graph partitioning algorithm, Boruvka’s algorithm, and a randomized algorithm for object graph clustering. Several points are innovative in our approach to clustering: (1) the randomized algorithm represents a new approach to the problem of clustering and is based on probabilistic combinatorics. (2) All of our algorithms can be used to cluster objects with multiple connectivity. (3) Currently applied partition-based clustering algorithms are based on Kruskal’s algorithm which always runs significantly slower than Prim’s and also uses considerably more storage. However in our implementation of clustering algorithms we achieved additional reduction in processing time of object graphs.

Original languageEnglish
Title of host publicationAdvances in Databases and Information Systems - 2nd East European Symposium, ADBIS 1998, Proceedings
EditorsWitold Litwin, Tadeusz Morzy, Gottfried Vossen
Place of PublicationBerlin
PublisherSpringer, Springer Nature
Pages236-247
Number of pages12
Volume1475
ISBN (Electronic)9783540683094
ISBN (Print)3540649247, 9783540649243
DOIs
Publication statusPublished - 1998
Event2nd East European Symposium on Advances in Databases and Information Systems, ADBIS 1998 - Poznan, Poland
Duration: 7 Sept 199810 Sept 1998

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1475
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other2nd East European Symposium on Advances in Databases and Information Systems, ADBIS 1998
Country/TerritoryPoland
CityPoznan
Period7/09/9810/09/98

Keywords

  • buffering
  • clustering
  • object database system
  • performance

Fingerprint

Dive into the research topics of 'Clustering techniques for minimizing object access time'. Together they form a unique fingerprint.

Cite this