A set of routines to operate on Hankel with Hankel block matrices stored in compact FFT-based form.
new.hbhmat(F, L = (N + 1) %/% 2,
wmask = NULL, fmask = NULL, weights = NULL,
circular = FALSE)
is.hbhmat(h)
hbhcols(h)
hbhrows(h)
hbhmatmul(hmat, v, transposed = FALSE)
array to construct the trajectory matrix for.
the window length.
special parameters for shaped SSA case (see ssa).
wmask and fmask are logical matrices, window and factor masks respectively.
weights is integer matrix which denotes hankel weights for array elements. If 'NULL',
parameters for simple rectangular 2D SSA case are used.
logical vector of one or two elements, describes field topology. 'TRUE' means circularity by a corresponding coordinate. If vector has only one element, this element will be used twice.
matrix to operate on.
logical, if 'TRUE' the multiplication is performed with the transposed matrix.
vector to multiply with.
Konstantin Usevich
Fast Fourier Transform provides a very efficient matrix-vector multiplication routine for Hankel with Hankel blocks matrices. See the paper in 'References' for the details of the algorithm.
Korobeynikov, A. (2010) Computation- and space-efficient implementation of SSA. Statistics and Its Interface, Vol. 3, No. 3, Pp. 257-268