Initializing the em algorithm in Gaussian mixture models with an unknown number of components

Volodymyr Melnykov, Igor Melnykov

Research output: Contribution to journalArticlepeer-review

68 Citations (Scopus)

Abstract

An approach is proposed for initializing the expectationmaximization (EM) algorithm in multivariate Gaussian mixture models with an unknown number of components. As the EM algorithm is often sensitive to the choice of the initial parameter vector, efficient initialization is an important preliminary process for the future convergence of the algorithm to the best local maximum of the likelihood function. We propose a strategy initializing mean vectors by choosing points with higher concentrations of neighbors and using a truncated normal distribution for the preliminary estimation of dispersion matrices. The suggested approach is illustrated on examples and compared with several other initialization methods.

Original languageEnglish
Pages (from-to)1381-1395
Number of pages15
JournalComputational Statistics and Data Analysis
Volume56
Issue number6
DOIs
Publication statusPublished - Jun 1 2012

Keywords

  • EM algorithm
  • Eigenvalue decomposition
  • Gaussian mixture model
  • Initialization
  • Truncated normal distribution

ASJC Scopus subject areas

  • Statistics and Probability
  • Computational Mathematics
  • Computational Theory and Mathematics
  • Applied Mathematics

Fingerprint Dive into the research topics of 'Initializing the em algorithm in Gaussian mixture models with an unknown number of components'. Together they form a unique fingerprint.

Cite this