×

zbMATH — the first resource for mathematics

On transformations and determinants of Wishart variables on symmetric cones. (English) Zbl 0890.60016
Summary: Let \(x\) and \(y\) be independent Wishart random variables on a simple Jordan algebra \(V\). If \(c\) is a given idempotent of \(V\), write \(x= x_1+ x_{12}+ x_0\) for the decomposition of \(x\) in \(V(c,1)\oplus V(c, 1/2)\oplus V(c,0)\) where \(V(c,\lambda)\) equals the set of \(v\) such that \(cv= \lambda v\). We compute \(E(\text{det}(ax+ by))\) and some generalizations of it. We give the joint distribution of \((x_1, x_{12}, y_0)\) where \(y_0= x_0- P(x_{12}) x^{-1}_1\) and \(P\) is the quadratic representation in \(V\). In statistics, if \(x\) is a real positive definite matrix divided into the blocks \(x_{11}\), \(x_{12}\), \(x_{21}\), \(x_{22}\), then \(y_0\) is equal to \(x_{22.1}= x_{22}- x_{21} x^{-1}_{11}x_{12}\). We also compute the joint distribution of the eigenvalues of \(x\). These results have been known only when \(V\) is the algebra of Hermitian matrices with entries in the real or the complex field. To obtain our results, we need to prove several new results on determinants in Jordan algebras. They include in particular extensions of some classical parts of linear algebra like Leibniz’s determinant formula or Schur’s complement.

MSC:
60E99 Distribution theory
62E10 Characterization and structure theory of statistical distributions
17C99 Jordan algebras (algebras, triples and pairs)
15A15 Determinants, permanents, traces, other special matrix functions
PDF BibTeX XML Cite
Full Text: DOI