v1.6.0rc3

another attempt to fix authors

v1.6.0rc2

fix authors

v1.6.0rc1

fix classifiers

v1.6.0rc0: Fix build (#169)

Modernize and fix build

v1.5.3 (Zenodo release)

No code change, this release is created to trigger Zenodo.

v1.5.2

What's Changed

• Update ci skripts and fix types #154
• bootstrap.resample now works with method="extended" when input is multi-dimensional #153

Full Changelog: v1.5.1...v1.5.2

v1.5.1

installation instructions in readme

v1.5.0

Changelog

v1.5.0-beta

Pre-release for the upcoming v1.5.0 release

First full-featured release for general production use

1.0.1 (August 23, 2020)

• Minor fix to allow building from source.

1.0.0 (August 22, 2020)

API Changes

• Bootstrap and jackknife generators resample.bootstrap.resample and resample.jackknife.resample are now exposed to compute replicates lazily.
• Jackknife functions have been split into their own namespace resample.jackknife.
• Empirical distribution helper functions moved to a resample.empirical namespace.
• Random number seeding is now done through using numpy generators rather than a global random state. As a result the minimum numpy version is now 1.17.
• Parametric bootstrap now estimates both parameters of the t distribution.
• Default confidence interval method changed from "percentile" to "bca".
• Empirical quantile function no longer performs interpolation between quantiles.

Enhancements

• Added bootstrap estimate of bias.
• Added bias_corrected function for jackknife and bootstrap, which computes the bias corrected estimates.
• Performance of jackknife computation was increased.

Bug fixes

• Removed incorrect implementation of Studentized bootstrap.

Deprecations

• Smoothing of bootstrap samples is no longer supported.
• Supremum norm and MISE functionals removed.

Other

• Benchmarks were added to track and compare performance of bootstrap and jackknife methods.