get.projections: Internal functions for MEMD
In PierreMasselot/Library--emdr: Tools for EMD-regression

get.projections creates projection directions through Hammersley sequence. sifting performs the sifting process to obtain one MIMF. compute.mean.enveloppe computes the mean enveloppe for sifting.

get.projections(ndirections, p)

sifting(x, tt, projmat, stopping, tol, max.iter)

compute.mean.enveloppe(hk, tt = 1:n, projmat = diag(p))

`ndirections`	Integer giving the number of projections.
`p`	Integer giving the number of variables to decompose.
`x`	A numeric matrix containing the multivariate signal. Each column corresponds to a variable.
`tt`	A numeric vector the same size of `row(x)` containing time indices for the data. Allows for irregularly sampled signals.
`projmat`	A `ndirection` x `p` giving the projection vectors to compute the enveloppe.
`stopping`	Character indicating the stopping criterion of the sifting process. When `stopping = "absmean"` (the default) the criterion of Rilling et al. (2003) based on the mean enveloppe is used. When `stopping = "S"`, the stopping criterion of Huang et al. (2003) based on the number of iteration is used.
`tol`	A numeric vector givving the tolerance for the stopping criterion. A vector of length 3 when `stopping = "absmean"` and a single value when `stopping = "S"`.
`max.iter`	Integer giving the maximum number of iterations for the sifting process.
`hk`	Matrix containing the current MIMF prototype.