Skip to content

scikit-posthocs: v0.18.0

Latest

Choose a tag to compare

@maximtrp maximtrp released this 13 Nov 16:20
07e4c18

What's Changed

  • Preserve wide block DataFrame treatment labels by @antonsoo in #109
  • Include tied statistics in Mack-Wolfe permutation probabilities by @antonsoo in #110
  • Preserve small Tamhane and Duncan adjusted probabilities by @antonsoo in #111
  • Fix Page and Durbin validation after missing-data filtering by @maximtrp in #113
  • Use analytic two-group Dunnett T3 probabilities and strengthen t-test example by @maximtrp in #114
  • Reverse NPM ranks without negating bounded integer observations by @antonsoo in #112
  • Reject invalid permutation counts in NPM, Hartley, and Mack-Wolfe tests by @maximtrp in #115
  • Preserve nonsignificant relationships in compact-letter fallback by @maximtrp in #116
  • Strengthen statistical test oracles and remove network-dependent collection by @maximtrp in #117
  • Preserve integer grouping-column names in Dunnett and Steel tests by @antonsoo in #119
  • Preserve explicit zero and empty DataFrame column names by @antonsoo in #120
  • Preserve Tukey HSD sample groups with duplicate row indices by @antonsoo in #118
  • Fix crossbars drawn out of order in critical difference diagrams by @Geochelone666 in #108
  • Harden crossbar-order and many-to-one input regressions by @maximtrp in #123
  • Align Tukey HSD categorical samples with output labels by @maximtrp in #124
  • Fix nullable numeric inputs on older pandas by @maximtrp in #125
  • Fuzz unusual global and omnibus inputs and fix discovered edge cases by @maximtrp in #126

New Contributors

Full Changelog: v0.17.1...v0.18.0