On the representation theorem for exchangeable arrays (Q1122865)
From MaRDI portal
![]() | This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use this page instead for the normal view: On the representation theorem for exchangeable arrays |
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | On the representation theorem for exchangeable arrays |
scientific article |
Statements
On the representation theorem for exchangeable arrays (English)
0 references
1989
0 references
An array of random variables (r.v.) \(\{X_{ij}\), i,j\(\geq 1\}\) is called separately exchangeable if its distribution is the same as that of \(\{X_{p_ iq_ j}\), i,j\(\geq 1\}\) for all permutations \((p_ i)\) and \((q_ j)\); and jointly exchangeable if it is identically distributed with \(\{X_{p_ ip_ j}\), i,j\(\geq 1\}\) for every permutation \((p_ i).\) It is known that separate exchangeability is equivalent to the existence of a representing function f and i.i.d. r.v. \(\{\alpha,\xi_ i,\eta_ j,\zeta_{ij}\}\) such that \(X_{ij}=f(\alpha,\xi_ i,\eta_ j,\zeta_{ij})\); a similar criterion holds for joint exchangeability. Straightforward proofs of a known criterion, and of a new one, for the nonuniqueness of a representing function are given.
0 references
exchangeable arrays of random variables
0 references
representing function
0 references
conditional independence
0 references
orthogonal expansions
0 references