test_op_surrogate

Notedocblock 
test_op_surrogate(data, method, n_surrogates; chart_choice, m=3, d=1, alpha=0.05, rng)

Compute a surrogate-data hypothesis test for ordinal patterns and return an OPTestResultSurrogate with the surrogate critical value, p-value, and reject decision.

Unlike test_op() and test_op_bootstrap(), which test against an i.i.d. null, the null hypothesis here is determined by the surrogate method: e.g. RandomFourier() tests against a stationary linear Gaussian process, and AAFT()/IAAFT() against a monotonic static transform of one — making this a nonlinearity test rather than a generic dependence test.

  • data: the time series.
  • method: a surrogate method from TimeseriesSurrogates.jl, e.g. RandomFourier(), AAFT(), IAAFT(), RandomShuffle().
  • n_surrogates: number of surrogate replications.
  • chart_choice: one of Persistence(), UpDownBalance(), RotationalAsymmetry(), UpDownScaling(), DistanceToWhiteNoise(), Shannon(), ShannonExtropy(). For Shannon and ShannonExtropy, the statistic is in the logarithm base of the chart, which must be larger than 1. Both default to base 2 in ComplexityMeasures.jl; use Shannon(base=exp(1)) for the natural logarithm used in the papers. Statistic and critical value are both in that base, so the test decision and the p-value do not depend on it.
  • m=3: length of the ordinal patterns.
  • d=1: delay between observations of a pattern.
  • alpha: significance level (default 0.05).
  • rng: random number generator used for the surrogate generation. Note: This function is provided as a package extension: it becomes available once TimeseriesSurrogates.jl is loaded, i.e. after using TimeseriesSurrogates.