spectral (version 1.0.1)

spec.lomb: Lomb-Scargle Periodigram

Description

The Lomb-Scargle periodigram represents an statistical estimator for the amplitude and phase for a given frequency.

Usage

spec.lomb(y = stop("Missing y-Value"), x = stop("Missing x-Value"),
  f = NULL, ofac = 4)

Arguments

y

data vector

x

sampling vector

f

frequeny vector

ofac

in case f=NULL this value controlls the amount of frequency oversampling.

Value

The spec.lomb function returns an object of the type lomb, which is a list containg the following parameters:

A

A vector with amplitude spectrum

f

corresponding frequency vector

phi

phase vector

x,y

original data

p

p-value as statistical measure

Details

A given time series does not need to be evenly sampled. This means a time series mainly consists of data pairs x and y, which store the data and the sampling position (e.g. in time). Additionally, this method enables the user to analyse the data with respect to a given frequency vector, which can be artifical dense.

ofac

If the user does not provide a corresponding frequency vector, the ofac parameter causes the function to estimate $$nf = ofac*length(x)/2$$ equidistant frequencies.

p-value

The p-value gives the probability, wheater the estimated amplitude is NOT significant. However, if p tends to zero the amplidutde is significant. The user must decide which maximum value is acceptable, until an amplitude is not valid.

References

A. Mathias, F. Grond, R. Guardans, D. Seese, M. Canela, H. H. Diebner, und G. Baiocchi, "Algorithms for spectral analysis of irregularly sampled time series", Journal of Statistical Software, 11(2), pp. 1--30, 2004.

J. D. Scargle, "Studies in astronomical time series analysis. II - Statistical aspects of spectral analysis of unevenly spaced data", The Astrophysical Journal, 263, pp. 835--853, 1982.

See Also

filter.lomb

Examples

# NOT RUN {
# create two sin-functions
x_orig <- seq(0,1,by=1e-2)
y_orig <- sin(10*2*pi*x_orig) + 0.3*sin(2*2*pi*x_orig)

# make a 10% gap
i <- round(length(x_orig)*0.2) : round(length(x_orig)*0.3)
x <- x_orig
y <- y_orig
x[i] <- NA
y[i] <- NA

# calculating the lomb periodogram
l <- spec.lomb(x = x, y = y)
# select a frequency range
m <- rbind(c(9,11))
# select and reconstruct the most significant component
l2 = filter.lomb(l, x_orig, filt=m)

# plot everything
par(mfrow=c(2,1),mar = c(4,4,2,4))
plot(x,y,"l", main = "Gapped signal")
lines(l2$x, l2$y,lty=2)
legend("bottomleft",c("gapped","10Hz component"),lty=c(1,2))

plot(l,main = "Spectrum")

# }