HDEst: Hurvich-Deo estimate of number of frequencies to use in a...

In wconstan/fractal: A Fractal Time Series Modeling and Analysis Package

Description

Estimates the number of frequencies to use in a periodogram regression estimate of the Hurst parameter H of a long memory time series. Based on estimated spectrum of time series.

Usage

HDEst(NFT, sdf, A=0.3, delta=6/7)

Arguments

NFT	number of points used in the Fourier transform to generate the spectrum.
sdf	estimate of the spectrum of the time series (power, NOT dB).
A	parameter A of Hurvich-Deo model. Default: 0.3 (recommended).
delta	parameter delta of Hurvich-Deo model. Default: 6/7 (recommended).

Value

estimated optimum number of frequencies "m" to use in a periodogram regression estimate of the Hurst parameter H.

References

C.M. Hurvich and R.S. Deo (1999), Plug-in Selection of the Number of Frequencies in Regression Estimates of the Memory Parameter of a Long Memory Time Series, J. Time Series Analysis, 20(3), 331–341.

See Also

hurstBlock.

Examples

S <- sapa::SDF(beamchaos)
HDEst(NFT=length(S),as.vector(S))

wconstan/fractal documentation built on May 4, 2019, 2:03 a.m.