XgenSimple: Generate multiple time series given simple VAR structure
In yymmhaha/PackPaper1: Package related to my first paper

Description Usage Arguments Details Value Examples

Generate vector autoregression time series with user-specified correlation structure

1	XgenSimple(p, A, x0, t, sigmax)

`p`	a number. Number of covariates x.
`A`	a list. Each element is a transition (lag coefficient) matrix. Caution: some specifications can make the series unstationary
`x0`	a list. Each element is a vector of length p, initial value of x. The length depends on the time lags (number of A).
`t`	a number. Length of series, including the initial lags.
`sigmax`	a number. Standard deviation of noise of x.

This function has been used as the most basic data generator. It is also used in the XgenCorr, XgenCorr2 and intervention functions. Mathematical details see the book.

a list of components

`Big A`	The expanded transition matrix for all lags
`X`	The generated multiple time series
`lag`	The lag for VAR process
`Characteristic roots`	Absolute value for 1/eigen(BigA), if any is smaller than 1, process is unstationary and series can not be generated

A <- list(A1 = toeplitz(c(0.4, rep(0, 9))),
         A2 = toeplitz(c(0.2, rep(0, 9))),
         A3 = toeplitz(c(0.1, rep(0, 9))))
x0 <- list(x1 = rep(0, 10), x2 = rep(0, 10), x3 = rep(0, 10))
sigmax <- 1
nseries <- 10
nsize <- 50

XgenSimple(p = nseries, A = A, x0 = x0, t = nsize, sigmax = sigmax)