Outlier detection in the multiple cluster setting using the minimum covariance determinant estimator

Johanna Hardin, David M Rocke

Research output: Contribution to journalArticle

86 Scopus citations

Abstract

Mahalanobis-type distances in which the shape matrix is derived from a consistent high-breakdown robust multivariate location and scale estimator can be used to find outlying points. Hardin and Rocke (http://www.cipic.ucdavis.edu/~dmrocke/preprints.html) developed a new method for identifying outliers in a one-cluster setting using an F distribution. We extend the method to the multiple cluster case which gives a robust clustering method in conjunction with an outlier identification method. We provide results of the F distribution method for multiple clusters which have different sizes and shapes.

Original languageEnglish (US)
Pages (from-to)625-638
Number of pages14
JournalComputational Statistics and Data Analysis
Volume44
Issue number4
DOIs
StatePublished - Jan 28 2004

Keywords

  • Minimum covariance determinant
  • Outlier detection
  • Robust clustering

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Statistics, Probability and Uncertainty
  • Electrical and Electronic Engineering
  • Computational Mathematics
  • Numerical Analysis
  • Statistics and Probability

Fingerprint Dive into the research topics of 'Outlier detection in the multiple cluster setting using the minimum covariance determinant estimator'. Together they form a unique fingerprint.

  • Cite this