powered by
Obtain generalized Schur complement
schur(M, x, y, z)
symmetric positive definite matrix
indices of M to calculate with (see below)
Calculates \(M_{xy} - M_{xz} M^{zz} M_{zy}\), which (if M is a Gaussian covariance matrix) is the covariance between x and y after conditioning on z.
y defaults to equal x, and z to be the complement of \(x \cup y\).