cdvc(y, X, V)

Fitting a variance component model with zero-mean responses y, and positive semi-definite matrices I, V[2], ..., V[end], and covariance σ2[1]*I+σ2[2]*V[2]+ ...+σ2[end]*V[end].

Arguments

• y: n-by-1 response vector.
• V: V[i] is the i-th psd matrix.

Keyword Arguments

• verbose: logical, verbose display.
• maxiter: maximum number of iterations.
• funtol: covergence tolerance for objective value.
• γ: staring value of variance components.

Output

• γ: fitted variance components.
• nter: number of iterations.
• logl: fitted log-likelihood.