R/RandomARMod_nlin1.R

In ChrisKust/rexpar: Simplicial Depth for Explosive Autoregressive Processes

RandomARMod_nlin1 <- function(nobs, arp, power = 1, start = 0, cont = "0", sd = 0.2)
{
  y <- numeric(nobs)
  y[1] <- start
  
  y <- switch(cont, 
      "0" = 
      {
        for(i in 2:length(y))
        {
          e <- rnorm(nobs) * sd
          y[i] <- y[i - 1] + arp * y[i - 1]^power + e[i]
        }
        y
      },
      "1" =
      {
        u1 <- rnorm(nobs) * 1
        u2 <- (2 + rnorm(nobs) * 1)
        p1 <- rpois(nobs, 5 / 100)
        e <- u1 + p1 * u2
        for(i in 2:length(y))
        {
          y[i] <- y[i - 1] + arp * y[i - 1]^power + e[i] - 0.007127848
        }
        y
      },
      "2" =
      {
        alpha <- 10
        beta <- - 3.665129
        u <- runif(nobs)
        e <- beta - alpha * log(( - log(u)))
        for(i in 2:length(y))
        {
          y[i] <- y[i - 1] + arp * y[i - 1]^power + e[i]
        }
        y
      },
      "3" =
      {
        alpha <- 1.928
        beta <- - 2
        gamma <- 10
        u <- runif(nobs)
        e <- beta + alpha * ( - log(u))^( - 1 / gamma)
        for(i in 2:length(y))
        {
          y[i] <- y[i - 1] + arp * y[i - 1]^power + e[i]
        }
        y
      },
      "4" =
      {
        alpha <- 1
        beta <- 1
        e <- rgamma(nobs, alpha, beta) - 0.6932
        for(i in 2:length(y))
        {
          y[i] <- y[i - 1] + arp * y[i - 1]^power + e[i]
        }
        y
      },
      "7" =
      {
        for(i in 2:length(y))
        {
          y[i] <- y[i - 1] + arp * y[i - 1]^power
        }
        y
      },
      stop("Insert a valid notion!")
  )
  }