Generate R-many conditionally independent AR(1) varying time series

generate_AR1_vary() generates R-many time series from an AR(1) varying processes. Each time series AR(1) coefficient is uniformly chosen from the range provided by the min and max arguments.

Usage

generate_AR1_vary(n, R = 1, min = 0.45, max = 0.6, burn = 50)

Arguments

n

A numeric vector that determines the length of the time series generated. Must contain R-many entries. Time series may be of different lengths.

R

An optional numeric scalar (default 1) that determines how many time series to generate. Each time series is stored as a column-vector in the list output.

min

An optional numeric scalar (default 0.45) that determines the minimum value the AR(1) coefficient could possibly take.

max

An optional numeric scalar (default 0.60) that determines the maximum value the AR(1) coefficient could possibly take.

burn

An optional numeric scalar (default 50) that determines the n.start argument of the arima.sim() function.

Value

a list object that contains the following fields:

ts_list	returns an R-long list each containing an (n[r] \(\times\) 1) matrix of the generated time series.
true_phi	returns a (1 \(\times\) R) matrix of true generated AR(1) coefficients.

Examples

## An AR(1) process with coefficient(s) selected from $U(min,max)$
## with different length time-series
R <- 5
n <- c(rep(2000, R - 2), rep(1000, R - 3))
burn <- 100
ts <- generate_AR1_vary(
  n = n,
  R = 5
)$ts_list
## `$ts_list` output generates an R-long list that each
## contains an (n[r] x 1) matrix

## Plot
## Create an empty plot
plot(
  x = c(),
  y = c(),
  xlim = c(0, 2000),
  ylim = range(ts),
  ylab = "",
  xlab = "time"
)
for (r in 1:3) {
  lines(ts[[r]][, 1], col = "blue")
}
for (r in 4:R) {
  lines(ts[[r]][, 1], col = "red")
}


## An AR(1) process with coefficient(s) selected from $U(min,max)$
## with same length time-series
phi <- c(0.5)
R <- 5
n <- rep(2000, R)
burn <- 100
ts <- generate_AR1_vary(
  n = n,
  R = 5
)$ts_list
## `$ts_list` output generates an R-long list that each
## contains an (n[r] x 1) matrix

## Plot
## Create an empty plot
plot(
  x = c(),
  y = c(),
  xlim = c(0, 2000),
  ylim = range(ts),
  ylab = "",
  xlab = "time"
)
for (r in 1:R) {
  lines(ts[[r]][, 1], col = "blue")
}



## Output is a list containing a (1000 x 5) matrix and a 5 long vector
## containing the true phi values randomly generated to create the
## R many time series.