[Date Index] [Thread Index] [Author Index]

Truncated inverse Wishart distribution

• To: mathgroup at smc.vnet.net
• Subject: [mg126584] Truncated inverse Wishart distribution
• From: paul <paulvonhippel at yahoo.com>
• Date: Mon, 21 May 2012 05:57:14 -0400 (EDT)
• Delivered-to: l-mathgroup@mail-archive0.wolfram.com

I would like to calculate the mean of a truncated inverse Wishart distribution? I.e., if B is an inverse Wishart variable, then I'd like to calculate E(B|B<W), where B<W means that W-B is positive definite.

I have the answer for the degenerate case where B and W are scalars -- that is, where B is an inverse chi-square variable and W is a scalar constant. The answer then is a ratio of gamma functions, and you can get it using the TruncatedDistribution function and the InverseChiSquareDistribution.

It seems to me I should be able to work up from the scalar answer to the matrix answer, but I'm not sure how. Any hints most appreciated.