test_sop_bp_bootstrap

Notedocblock

test_sop_bp_bootstrap(data, n_boot, w; chart_choice, refinement, alpha, block_size)

Compute a bootstrap Box-Pierce type hypothesis test for spatial ordinal patterns (SOPs), aggregating the squared chart statistics over all delay combinations (d1, d2) ∈ {1,…,w}², and return a SOPBPTestResultBoot with the bootstrap critical value, p-value, and reject decision.

No asymptotic theory is available for the BP-SOP statistic, so — unlike the single-delay test_sop — this bootstrap test is the only classical test provided for it.

• data: the 2D image (M × N matrix).
• n_boot: number of bootstrap replications.
• w: window size; all (d1, d2) ∈ {1,…,w}² are included in the BP sum.
• chart_choice: one of TauHat(), KappaHat(), TauTilde(), KappaTilde().
• refinement: OrdinaryType() for the classical SOP classification, or a RefinedType instance.
• alpha: significance level (default 0.05).
• block_size: set > 1 for a 2D block bootstrap that preserves spatial dependencies.