ARp: Autoregressive model of order _p_

In TycheLab/CoSMoS: Complete Stochastic Modelling Solution

Description

Generates time series from an Autoregressive model of order p.

Usage

ARp(margdist, margarg, acsvalue, actfpara, n, p = NULL, p0 = 0)

Arguments

margdist	target marginal distribution
margarg	list of marginal distribution arguments
acsvalue	target auto-correlation structure (from lag 0)
actfpara	auto-correlation structure transformation parameters
n	number of values
p	integer - model order (if NULL - limits maximum model order according to auto-correlation structure values)
p0	probability zero

Examples

library(CoSMoS)

## choose the marginal distribution as Pareto type II with corresponding parameters
dist <- 'paretoII'
distarg <- list(scale = 1, shape = .3)
p0 <- .5

## estimate rho 'x' and 'z' points using ACTI
pnts <- actpnts(margdist = dist, margarg = distarg, p0 = p0)

## fit ACTF
fit <- fitactf(pnts)

## define target auto-correlation structure and model order
order <- 1000
acsvalue <- acs(id = 'weibull', t = 0:order, scale = 10, shape = .75)

## limit ACS lag (recomended)
system.time(val <- ARp(margdist = dist,
                       margarg = distarg,
                       acsvalue = acsvalue,
                       actfpara = fit,
                       n = 5000,
                       p0 = p0))

## order w/o limit
system.time(val <- ARp(margdist = dist,
                       margarg = distarg,
                       acsvalue = acsvalue,
                       actfpara = fit,
                       n = 5000,
                       p = order,
                       p0 = p0))


## see the result
ggplot() +
  geom_col(aes(x = seq_along(val),
               y = val)) +
  labs(x = '',
       y = 'value') +
  theme_classic()

TycheLab/CoSMoS documentation built on June 6, 2021, 2:35 a.m.