tsc_dep_var: Timescale-dependent variance estimation

In krehfeld/nest: Non-equally spaced time series analysis

Description

Estimate the timescale-dependent variance of a time series.

Usage

tsc_dep_var(timser, tsc.in, min.res =
                 min(mean(diff(index(timser))), floor(tsc.in[1]/2) -
                 1), start.val = start(timser), end.val = end(timser),
                 pval = 0.1, detrend = FALSE)

Arguments

timser	Input time series. Can be a zoo, ts or vector object.
tsc.in	Vector of two timescales, c(tsc1,tsc2).
min.res	Minimal resolution of the time series (if this is not met, NAs are returned).
start.val	Starting value of the time series window.
end.val	End value of the time series window.
pval	pval gives the p-value for the confidence intervals of the variance estimate.
detrend	logical argument, if set to TRUE the timesereis is linearly detrended prior to spectral analysis

Value

A list of

std	Standard deviation
var.tsc	timescale-dependent variance estimated
var.tot	total variance of the time series in the window
dof	estimated degrees of freedom of the variance estimate
ts.used	used time series
var.ci	list of lo and up, lower and upper confidence level for the variance estimate
spec.win	spectrum object

Author(s)

Kira Rehfeld, with contributions from Thom Laepple

See Also

var_ci

Examples

library(scales)
## Generate one gamma-distributed and one regular time axis
tx<-generate_t(dt=1,tmin=0,tmax=250,method="gamma")
ty<-generate_t(dt=1,tmin=0,tmax=250,method="linear")
## Simulate one coupled AR1 process (see reference for details)
Proc<-car(tx,ty,coupl_strength=0.5,phi=0.5,lag=0,nsur=1)
## Bind the results to zoo time series
x<-zoo(Proc$x,order.by=tx)
y<-zoo(Proc$y,order.by=ty)


tsc_dep_var(y,tsc.in=c(30,100)) 

krehfeld/nest documentation built on May 28, 2019, 12:33 a.m.