phdu.stats.rtopy package

Submodules

phdu.stats.rtopy.resample module

Adaptation of resample R package: https://cran.r-project.org/web/packages/resample/resample.pdf Based on the article: https://arxiv.org/abs/1411.5279

phdu.stats.rtopy.resample.CI_bootstrap(x, y=None, stat='mean', method='bootstrapT', expand=True, alpha=0.05, probs=None, alternative='two-sided', return_stats=False, **kwargs)[source]

Check for the stats. Unbiased, functional statistics such as the mean should have zero bootstrap bias. If not, increase N. NOTE: s2 will have non-zero bootstrap bias because it is not functional.

Attrs:

  • method: [“t”, “percentile”, “bca”, “bootstrapT”].

  • alternative: [“two-sided”, “less”, “greater”].

phdu.stats.rtopy.resample.bootstrap(x, method, y=None, stat='mean', N=100000, seed=0, block_size=1000, **kwargs)[source]
Attrs:

  • method: [“t”, “percentile”, “bca”, “bootstrapT”]

Returns: bootstrap object (R)

phdu.stats.rtopy.resample.permutation(x, y=None, stat='mean', N=100000, seed=0, block_size=1000, **kwargs)[source]

Returns: permutationTest object (R) and the results of the test.

The implementation in numba is faster and includes pairing option, which this one does not.

phdu.stats.rtopy.resample.stat_computer(obj_name, nsamples=1)[source]

Module contents