Gaussian.rotate_matrix(R1, R2, inv1=None, logdet1=None, inv2=None, logdet2=None, Q=None)[source]

The vector is reshaped into a matrix by stacking the row vectors.

Computes R1*X*R2’, which is identical to kron(R1,R2)*x (??)

Note that this is slightly different from the standard Kronecker product definition because Numpy stacks row vectors instead of column vectors.

R1 : ndarray

A matrix from the left

R2 : ndarray

A matrix from the right