Probabilistic validation approach for clustering

M. Har-even, V. L. Brailovsky*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

The suggested approach combines the phases of cluster validity and cluster tendency inside the scope of the clustering algorithm. The algorithm is based on a probabilistic approach and is invariant to the scaling of features. The result is an efficient algorithm whose performance is demonstrated on real and synthetic data.

Original languageEnglish
Pages (from-to)1189-1196
Number of pages8
JournalPattern Recognition Letters
Volume16
Issue number11
DOIs
StatePublished - Nov 1995

Keywords

  • Cluster analysis
  • Probabilistic validation
  • Projection pursuit
  • Simulating annealing
  • Unsupervised hierarchical clustering

Fingerprint

Dive into the research topics of 'Probabilistic validation approach for clustering'. Together they form a unique fingerprint.

Cite this