The non-central chi-square distribution with $\nu$ degrees of freedom has a density which can be expressed as $$ f(x) = \sum_{i=0}^{\infty} c_i f_{\nu + 2i}(x), $$ where $c_i$ is a function of the non-centrality parameter and $i$, and where $f_{j}(x)$ is the density of a chi-square distribution with $j$ degrees of freedom. (The non-central chi-squared distribution is a Poisson weighted mixture of central chi-squared distributions).

My question is whether this property extends to the multivariate case, the non-central Wishart distribution: can we express the non-central Wishart as a weighted mixture of central Wishart distributions?

(My guess is not, except in the scalar case; if it does hold, however, and the $c_i$ are easy to compute, I would like to use this property to compute the moments of the non-central Wishart.)

  • $\begingroup$ It looks like Theorem 1 of Li and Geng provides something like what I would need, for the case where the non-centrality matrix is rank 1. $\endgroup$ – Steven Pav Mar 23 '18 at 4:53

arXiv:1512.08159 shows that the distribution of the scalar Schur complement in a noncentral Wishart matrix is a mixture of central chi-square distributions with different degrees of freedom. For the case of a rank-1 noncentrality matrix, the weights of the mixture representation arise from a noncentral beta mixture of Poisson distributions.

If the goal is to calculate the moments of the noncentral Wishart distribution, the graph representation of arXiv:0912.0577 seems convenient. (See also the tutorial On the moment formulas for the noncentral Wishart distributions.)


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.