Skip to contents

Changelog

Source: NEWS.md

riemstats 0.2.0

Major Changes

Permutation Test Implementation

  • BREAKING CHANGE: Replaced parametric bootstrap with permutation test in riem_anova()
    • Function parameter renamed: den → nperm (default changed from 5 to 1000)
    • Old function one_bootstrap() removed and replaced with one_permutation()
    • Users must update existing code to use the new parameter name

Benefits of permutation test approach: - Faster computation (no synthetic data generation required) - More stable p-values (no parameter estimation variability) - Exact test under null hypothesis (when group labels are exchangeable) - Simpler implementation (direct shuffling of observations)

Motivation: - Asymptotic tests showed inflation issues - Bootstrap was computationally expensive and under-conservative - Permutation test is more natural for comparing sub-populations

See GitHub issue #61 and PR #62 for detailed discussion and implementation.

Documentation

  • Updated riem_anova() documentation with permutation test description
  • Added one_permutation() function documentation
  • Updated vignette with working permutation test examples
  • Reduced vignette permutation count to 100 for faster builds

Testing

  • Updated test suite for permutation approach
  • All 126 tests passing
  • Fixed CSample constructor issues in tests

Note on Validation

This release is pending comprehensive validation through simulation studies (EXP-MP-062). Initial testing shows correct behavior, but full Type I error and power analysis is ongoing.

riemstats 0.1.0

Dependencies

  • Depends on R (>= 4.3.0)
  • Imports the following packages:
    • Matrix
    • methods
    • expm
    • R6
    • purrr
    • MASS
    • furrr
  • Suggests the following packages for testing and documentation:
    • testthat (>= 3.0.0)
    • knitr
    • rmarkdown

Miscellaneous

  • Added a LICENSE file with the MIT license.
  • Maintainer: Nicolas Escobar