location (real
scale (positive-definite real
matrix) In statistics, the matrix normal distribution or matrix Gaussian distribution is a probability distribution that is a generalization of the multivariate normal distribution to matrix-valued random variables.
The probability density function for the random matrix X (n × p) that follows the matrix normal distribution
denotes trace and M is n × p, U is n × n and V is p × p, and the density is understood as the probability density function with respect to the standard Lebesgue measure in
, i.e.: the measure corresponding to integration with respect to
The matrix normal is related to the multivariate normal distribution in the following way: if and only if where
denotes the Kronecker product and
denotes the vectorization of
The equivalence between the above matrix normal and multivariate normal density functions can be shown using several properties of the trace and Kronecker product, as follows.
We start with the argument of the exponent of the matrix normal PDF: which is the argument of the exponent of the multivariate normal PDF with respect to Lebesgue measure in
The proof is completed by using the determinant property:
denotes trace.
More generally, for appropriately dimensioned matrices A,B,C: Transpose transform: Linear transform: let D (r-by-n), be of full rank r ≤ n and C (p-by-s), be of full rank s ≤ p, then: Let's imagine a sample of n independent p-dimensional random variables identically distributed according to a multivariate normal distribution: When defining the n × p matrix
for which the ith row is
, we obtain: where each row of
is the n × n identity matrix, that is the rows are independent, and
Given k matrices, each of size n × p, denoted
, which we assume have been sampled i.i.d.
from a matrix normal distribution, the maximum likelihood estimate of the parameters can be obtained by maximizing: The solution for the mean has a closed form, namely but the covariance parameters do not.
However, these parameters can be iteratively maximized by zero-ing their gradients at: and See for example [3] and references therein.
The covariance parameters are non-identifiable in the sense that for any scale factor, s>0, we have: Sampling from the matrix normal distribution is a special case of the sampling procedure for the multivariate normal distribution.
be an n by p matrix of np independent samples from the standard normal distribution, so that Then let so that where A and B can be chosen by Cholesky decomposition or a similar matrix square root operation.
Dawid (1981) provides a discussion of the relation of the matrix-valued normal distribution to other distributions, including the Wishart distribution, inverse-Wishart distribution and matrix t-distribution, but uses different notation from that employed here.