Identifiability of Mixtures

G. M. Tallis, Peter Chesson

Research output: Contribution to journalArticle

18 Scopus citations

Abstract

Let F(x, 0) be a family of distribution functions indexed by If G(0) is a distribution function on 12, H(x) = fQF(x, 0) dG(B) is a mixture with respect to G. If there is a unique G yielding H, the mixture is said to be identifiable. This paper summarises some known results related to identifiability of special types of mixtures and then discusses the general problem of identifiability in terms of mappings. Some new results follow for mappings with special features.

Original languageEnglish (US)
Pages (from-to)339-348
Number of pages10
JournalJournal of the Australian Mathematical Society
Volume32
Issue number3
DOIs
StatePublished - 1982
Externally publishedYes

Keywords

  • bounded inverse
  • identifiability
  • kernel
  • linear independence
  • mixture
  • strong independence
  • unbounded inverse

ASJC Scopus subject areas

  • Mathematics(all)

Fingerprint Dive into the research topics of 'Identifiability of Mixtures'. Together they form a unique fingerprint.

  • Cite this