simulateFluctuations: Simulate time series of fluctuations

In MJAlexander/distortr: Temporal smoothing methods for demographic time series

Description

Simulate a time series of fluctuations based on specified underlying process. Options are AR(1), ARMA(1,1), P-splines or GPR.

Usage

simulateFluctuations(nyears, prop.sample = 1, method, params, obs.err = T,
  sigma.y = NULL, seed = 123)

Arguments

nyears	number of years of observation period
prop.sample	proportion of total years to 'observe'. Default is 1.
method	string to indicate type of method from which to simulate. Options are: "ar", "arma", "splines", "gp".
params	list of parameters needed to simulate from method specified. Each method takes different parameter arguments. See documentation for each process for more information. AR: specify nyears, rho, sigma ARMA: specify nyears, phi, theta, sigma.ar Splines: specify x.i, degree GP: specify covariance function (sqexp or matern). For sqexp specify nyears, tau, l. For matern specify nyears, tau, smoothness and range.
obs.err	whether or not to include observational error. Default is TRUE.
sigma.y	value of sd of observational error, if included
seed	value of random seed

Value

A data frame of observation times and y-values

See Also

GetAR, GetARMA, GetPSplines, GetGP

Examples

nyears <- 100
prop.sample <- 0.7
obs.err <- TRUE
sigma.y <- 0.5
seed <- 123
method <- "ar"
params <- list(rho = 0.5, sigma = 1, ystart = NULL, eps0.t = NULL)
res <- simulateFluctuations(nyears, prop.sample, method, params, obs.err, sigma.y)
ggplot(data = res, aes(x = t, y = y)) + geom_line() + geom_point() + theme_bw()

MJAlexander/distortr documentation built on July 17, 2020, 4:06 p.m.