XgenCorr: Generate blockly correlated VAR time series (type 1)

In yymmhaha/PackPaper1: Package related to my first paper

Description

Generate vector autoregression time series with 3 blockly correlated structure

Usage

XgenCorr(p.block, x0.block, A.block, t, sigmax)

Arguments

p.block	a list of 3 numbers. Number of covariates for each block. Default is 3, 3, 4 so they form a 10 series X.
x0.block	a list of 3 sublists. Each sublist contains vectors as the initial values
A.block	a list of 3 sublists. Each sublist contains AR coefficient matrices.
t	a number. Length of series, including the initial lags.
sigmax	a number. Standard deviation of noise of x.

Details

This function is based on XgenSimple. 3 blocks, hence need specification for each block, and arguments p.block, x0.block, A.block should have length of 3. It is recommended to use coef.mat to generate the AR coefficient matrices.

Value

a matrix

bigX	The combined matrix of 3 individual submatrices

Examples

A11 <- coef.mat(dimension = 3, ondiag = 0.4, offdiag = 0.15)
A12 <- coef.mat(dimension = 3, ondiag = 0.1, offdiag = 0.07)
A13 <- coef.mat(dimension = 3, ondiag = 0.1, offdiag = 0)

A21 <- coef.mat(dimension = 3, ondiag = 0.4, offdiag = 0.15)
A22 <- coef.mat(dimension = 3, ondiag = 0.1, offdiag = 0.05)
A23 <- coef.mat(dimension = 3, ondiag = 0.1, offdiag = 0)

A31 <- coef.mat(dimension = 4, ondiag = 0.4, offdiag = 0.15)
A32 <- coef.mat(dimension = 4, ondiag = 0.1, offdiag = 0.05)
A33 <- coef.mat(dimension = 4, ondiag = 0.1, offdiag = 0)

A.block <- list(block1 = list(A1 = A11, A2 = A12, A3 = A13),
                block2 = list(A1 = A21, A2 = A22, A3 = A23),
                block3 = list(A1 = A31, A2 = A32, A3 = A33))
p.block <- list(p1 = 3, p2 = 3, p3 = 4)


x11 <- x12 <- x13 <- rep(0, 3)
x21 <- x22 <- x23 <- rep(0, 3)
x31 <- x32 <- x33 <- rep(0, 4)

x0.block <- list(x10 = list(x1 = x11, x2 = x12, x3 = x13),
                 x20 = list(x1 = x21, x2 = x22, x3 = x23),
                 x30 = list(x1 = x31, x2 = x32, x3 = x33))

yymmhaha/PackPaper1 documentation built on May 24, 2019, 8:55 a.m.