Characterization theorems for some classes of covariance functions associated to vector valued random fields

Emilio Porcu, Viktor Zastavnyi

Research output: Contribution to journalArticlepeer-review

40 Scopus citations

Abstract

We characterize some important classes of cross-covariance functions associated to vector valued random fields based on latent dimensions. We also give some results for mixture based models that allow for the construction of new cross-covariance models. In particular, we give a criterion for the permissibility of quasi-arithmetic operators in order to construct valid cross covariances.

Original languageBritish English
Pages (from-to)1293-1301
Number of pages9
JournalJournal of Multivariate Analysis
Volume102
Issue number9
DOIs
StatePublished - Oct 2011

Keywords

  • Cross-covariance functions
  • Exponentially convex functions
  • Latent dimensions
  • Multivariate laplace transforms
  • Quasi-arithmetic operators
  • Vector valued random fields

Fingerprint

Dive into the research topics of 'Characterization theorems for some classes of covariance functions associated to vector valued random fields'. Together they form a unique fingerprint.

Cite this