tseriesEntropy (version 0.7-2)

surrogate.SA: Surrogate Time Series Through Simulated Annealing

Description

Starting from a time series x given as input, the function generates surrogate series through Simulated Annealing. Each surrogate series is a constrained random permutation having the same autocorrelation function (up to nlag lags) of the original series x. The surrogates can be used for testing for non linearity in time series.

Usage

surrogate.SA(x, nlag, nsurr, Te = 0.0015, RT = 0.9, eps.SA = 0.05, nsuccmax = 30,
 nmax = 300, che = 1e+05)

Value

A list with the following elements:

surr

a matrix with N rows and nsurr columns, in each column is stored a surrogate.

call

contains the call to the routine.

Arguments

x

a univariate numeric time series object or a numeric vector.

nlag

minimization is performed w.r.t. to the first nlag lags.

nsurr

number of surrogates.

Te

starting value for the temperature.

RT

reduction factor for the temperature Te.

eps.SA

target tolerance.

nsuccmax

Te is decreased after nsuccmax*N successes.

nmax

Te is decreased after nmax*N successes.

che

after che*2N global iterations the algorithm starts again.

Author

Simone Giannerini<[e-mail]>

Details

Let N be the length of the series x. Sensible (N-dependent) defaults are derived for the parameters of the algorithm, there should not be the need to change them. In case, the user could try increasing eps.SA.

References

Giannerini S., Maasoumi E., Bee Dagum E., (2015), Entropy testing for nonlinear serial dependence in time series, Biometrika, 102(3), 661--675 tools:::Rd_expr_doi("10.1093/biomet/asv007").

Schreiber T., Schmitz A.,(2000) Surrogate time series. Physica D, 142(3-4), 346--382.

See Also

See Also Trho.test.SA, surrogate.AR, Trho.test.AR.

Examples

set.seed(1345)
# Generates a AR(1) series
x      <- arima.sim(n=120, model = list(ar=0.8));
x.surr <- surrogate.SA(x, nlag=10, nsurr=3);
plot.ts(x.surr$surr,col=4);


## Check that the surrogates have the same ACF of x
corig <- acf(x,10,plot=FALSE)$acf[,,1];
csurr <- acf(x.surr$surr[,1],10,plot=FALSE)$acf[,,1];
round(cbind(corig,csurr,"abs(difference)"=abs(corig-csurr)),3)