Constructs Q = lambda * (D_q' D_q) kronecker I_d, with optional ridge.
make_random_walk_precision(K, d, q = 1, lambda = 1, ridge = 0, sparse = FALSE)
Arguments
- K
integer number of time points.
- d
integer dimension of the process at each time.
- q
integer order of the random walk (q>=1).
- lambda
nonnegative scalar precision multiplier.
- ridge
nonnegative scalar ridge added to the 1D precision.
- sparse
logical; return a sparse Matrix object if TRUE.
Value
(dK) x (dK) precision matrix.