innovationVariance: Nonparametric estimate of the innovation variance

In ltsa: Linear Time Series Analysis

Description

The innovation variance is estimated using a high order AR approximation determined by the AIC or by using Kolmogoroff's formula with a smoothed periodogram. Default is AR.

Usage

innovationVariance(z, method = c("AR", "Kolmogoroff"), ...)

Arguments

z	time series
method	Default "AR". Set to "Kolmogoroff" for non-parametric periodogram estimate.
...	optional arguments that are passed to spec.pgram()

Value

the innovation variance

Author(s)

A. I. McLeod

See Also

exactLoglikelihood, PredictionVariance,

Examples

z<-sunspot.year
#fitting high-order AR
innovationVariance(z)
#using periodogram
innovationVariance(z, method="Kolmogoroff")
#using smoothed periodogram
innovationVariance(z, method="Kolmogoroff", span=c(3, 3))
#the plot argument for spec.pgram() works too
innovationVariance(z, method="Kolmogoroff", span=c(3, 3), plot=TRUE)

Example output

[1] 222.5218
[1] 161.5388
[1] 260.5773
[1] 260.5773

ltsa documentation built on May 2, 2019, 4:01 a.m.