Greetings,
I am working on a research problem for which I would like to use a
multivariate student t distribution (MVT). As part of this work, I
would like to calculate the conditional expectation of a vector of
random variables, partitioned as follows:
Let z be an n x 1 vector of random variables distributed MVT. Now
partition z = (z1,z2), such that z1 is n1 x 1 and z2 is n2 x 1 , with
n1 + n2 = n.
I would like to find the formula for the conditional expectation of
z1 given z2, i.e., E(z1| z2,R), where R is a given covariance matrix.
Can anyone point me towards a source in the literature? I believe
that problems of this sorts often occur in the Bayesian analysis of
linear regression models.
Thanks,
-Mark
|