ptest (version 1.0-8)

pgram: Periodogram computation

Description

The periodogram is computed.

Usage

pgram(z, fr = "default", method = c("periodogram", "regression"))

Arguments

z

time series vector of length n, say.

fr

use "default" for usual Fourier frequencies, 1/n, ..., floor(n/2)/n. Set fr = N, to evaluate the periodogram at the Fourier frequencies corresponding to a time series of length N. Finally set fr to any desired set of frequencies. Note frequencies are in cycles per unit time sometimes called temoral frequency to distinguish from angular frequency. Both are widely used in time series.

method

either periodogram or regression

Value

Periodogram evaluated at the Fourier frequencies or R-square.

Details

Uses FFT. So if the length of z is a highly composite number, the computation is very efficient. Otherwise the usual DFT is used.

Examples

z<-sunspot.year
n<-length(z) 
I<-pgram(z)
f<-I[,1]
I <- I[,2]
plot(f, I, xlab="f", ylab="f", type="l") 
title(main="Periodogram for Annual Sunpots, 1700-1988") 
#
z<-c(0.42, 0.89, 1.44, 1.98, 2.21, 2.04, 0.82, 0.62, 0.56, 0.8, 1.33)
fr <- (1:50)/101
pgram(z)
pgram(z, fr=101)
pgram(z, fr=fr)
pgram(z, method="regression")
pgram(z, method="regression", fr=101)
pgram(z, method="regression", fr=fr)