linproc: Generate a linear process from

In freqdom: Frequency Domain Analysis for Multivariate Time Series

Description

Compute

Y_t = ∑_{k=-q}^p A_k X_{t-k}

where X_t is a stationary multivariate time series and (A_k)_{-q ≤q k ≤q p} is a filter.

Usage

linproc(X, A, noise = NULL)

Arguments

X	process process
A	time-domain operator series
noise	function taking dimension D and returning D-dimentional vector

Value

Multivariate linear process

See Also

speclagreg

Examples

d = 2
n = 100
X = rar(n,d=d)

OP = array(0,c(d,d,2))
OP[,,1] = 2 * diag(d:1)/d
OP[,,2] = 1.5 * diag(d:1)/d
A = timedom(OP, 0:1)

Y = linproc(X,A,noise=rnorm)

freqdom documentation built on May 2, 2019, 5:55 p.m.