surrogate_ensemble: Generates surrogate series for 'ts', which is a numeric...

Description Usage Arguments Details

Description

Implemented methods are: "random" (Randomly shuffle data) "phase" (Theiler's phase randomization - constrained) "ebisuzaki" (Another form (identical?) of phase randomisation) "aaft" (Theiler's AAFT - constrained) "ce" (Davies and Harte's Circulant Embedding - nonconstrained) "dh" (Davison and Hinkleys phase and amplitude randomization - nonconstrained)

Usage

1
2
surrogate_ensemble(ts, surrogate.method = "random", n.surrogates = 1,
  print.to.console = F)

Arguments

ts

Time series for which to generate surrogate time series. A vector.

surrogate.method

String indicating the surrogate time series generation method.

n.surrogates

The number of surrogates to generate.

print.to.console

Display run info?

Details

The null hypotheses for these methods are (information about the methods of the 'fractal' package is mostly paraphrased or directly copied from its R-documentation): "random": The data represents white noise. "phase": The original data can be reproduced by a linear Gaussian process. The phase at each frequency of the Fourier transform of the original time series is randomised to be uniformly distributed on [0, 2*PI]. Periodograms of the surrogate and the original time series are the same. "ebisuzaki": "aaft: Create normally distributed white noise, rank-ordered according to the original time series. The white noise is then phase randomised, and its rank is calculated. The surrogate is then produced by rank ordering the original time series according to the new rank. Surrogate preserves the amplitude distribution (histogram) of the original time series. "ce": Circulant embedding. Generate surrogate whose estimated SDF (periodogram) is not constrained to be the same as for the original time series. Unconstrained. "dh": Davison-Hinkley surrogates. Randomises both the phases and the amplitudes in the frequency domain. The surrogate is then produced by inverting back to the time domain. Unconstrained.


kahaaga/tstools documentation built on May 24, 2019, 5:01 a.m.