powered by
Compute the mean of prod(x)^power when x follows T_dof(mu,sigma) distribution (dof= -1 for multivariate Normal).
eprod(m, S, power = 1, dof = -1)
Expectation of the above-mentioned product
Location parameter
Scale matrix. For multivariate T with dof>2 the covariance is S*dof/(dof-2). For the multivariate Normal the covariance is S.
Power that the product is raised to
Degrees of freedom of the multivariate T. Set to -1 for the multivariate Normal.
John Cook
The calculation is based on the computationally efficient approach by Kan (2008).
Kan R. From moments of sum to moments of product. Journal of Multivariate Analysis 99 (2008), 542-554.
#Check easy independence case m <- c(0,3); S <- matrix(c(2,0,0,1),ncol=2) eprod(m, S, power=2) (m[1]^2+S[1][1])*(m[2]^2+S[2][2])
Run the code above in your browser using DataLab