cep.mtm: Multitaper Estimation of Cepstral Coefficients and the...

In CepLDA: Discriminant Analysis of Time Series in the Presence of Within-Group Spectral Variability

Description

Returns multitaper estimated cepstra coefficients and log-spectrum for univariate time series.

Usage

cep.mtm(x,nw,k)

Arguments

x	Univariate time series of length N.
nw	Width of tapers. Default is set to 4
k	Number of tapers. Default is set to 7

Value

a list with 4 elements

quef	Quefencies.
cep	Raw cepstra coefficients from 0:(N-1)
freq	Frequencies between 0 and 1
lspec	Log-spectrum.

Author(s)

Robert Krafty <[e-mail]>

Examples

## simulate a time series
N = 500 #length of each series
dat <- r.cond.ar2(N=N,nj=1,r.phi1=c(.01,.7),r.phi2=c(-.12,-.06),r.sig2=c(.3,3))$X

## Fit multiaper
cep <- cep.mtm(dat)

## Plot the cepstral coefficients
plot(cep$quef, cep$cep)

## Plot the log spectrum
plot(cep$freq, cep$lspec, type="l")

CepLDA documentation built on May 1, 2019, 6:50 p.m.