Fairness.jl Versions Save

Julia Toolkit with fairness metrics and bias mitigation algorithms

v0.3.2

2 years ago

Fairness v0.3.2

Diff since v0.3.1

Merged pull requests:

  • CompatHelper: add new compat entry for "Downloads" at version "1.4" (#62) (@github-actions[bot])
  • CompatHelper: bump compat for "MLJBase" to "0.18" (#63) (@github-actions[bot])
  • Added Calibrated Equalised Odds Wrapper. (#64) (@samurai-kant)
  • CompatHelper: bump compat for "MLJModelInterface" to "1.0" (#65) (@github-actions[bot])
  • CompatHelper: bump compat for "DataFrames" to "1.0" (#66) (@github-actions[bot])
  • CompatHelper: bump compat for "CategoricalArrays" to "0.10" (#67) (@github-actions[bot])

v0.3.1

3 years ago

Fairness v0.3.1

Diff since v0.3.0

Merged pull requests:

  • CompatHelper: bump compat for "MLJBase" to "0.17" (#60) (@github-actions[bot])

v0.3.0

3 years ago

Fairness v0.3.0

Diff since v0.2.0

v0.2.0

3 years ago

Fairness v0.2.0

Diff since v0.1.3

Merged pull requests:

  • Synthetic data (#42) (@pfistfl)
  • Update README.md (#43) (@vollmersj)
  • Synthetic data (#44) (@pfistfl)
  • CompatHelper: bump compat for "GLPK" to "0.14" (#45) (@github-actions[bot])
  • fix errror in subgroupData (#46) (@pfistfl)
  • Change postprocessing structs subtyping (#47) (@jademackay)
  • Update fair-split branch (#49) (@ashryaagr)

v0.1.3

3 years ago

Fairness v0.1.3

  • Bank Marketing, Communities Crime, Student Performance Datasets
  • Benchmarking pipeline and Disparate Impact metric
  • fix for mach() deprecation in MLJ
  • Windows support and appveyor testing
  • MacOS testing using travis

Diff since v0.1.2

Merged pull requests:

  • added Literate example from next journal (#38) (@vollmersj)
  • CompatHelper: bump compat for "MLJModels" to "0.12" (#39) (@github-actions[bot])
  • CompatHelper: bump compat for "MLJBase" to "0.15" (#40) (@github-actions[bot])
  • CompatHelper: add new compat entry for "ZipFile" at version "0.9" (#41) (@github-actions[bot])

v0.1.2

3 years ago

Fairness v0.1.2

  • First Inprocessing Algorithm PenaltyWrapper Added.
  • Performance enhancements through:
    • Fairness Tensor made immutable
    • Fairness Algorithm Wrappers made immutable
  • LinProgWrapper modified to support immutable FairTensor
  • setindex for FairTensor added

Diff since v0.1.1

v0.1.1

3 years ago

Fairness v0.1.1

First Official Release of Fairness.jl

Changes

  • Fixed LinProgWrapper algorithm (The generalized Equalized Odds)
  • Added accuracy metric
  • Silenced warnings in Adult Dataset macro
  • Bug fix in Metric Wrapper
  • Disparity integrated with MLJ.evaluate

v0.1.0

3 years ago

Fairness v0.1.0

First Release of Fairness.jl

It has Datasets, Metrics, Algorithms, and integration with MLJ. The package has been extensively documented and has Comprehensive test coverage.

Merged pull requests:

  • Update calcmetrics.jl (#4) (@vollmersj)
  • Add COMPAS and Adult dataset. (#6) (@Faldict)
  • Added clean! and target scitypes for fairness algorithm-wrappers (#7) (@ashryaagr)
  • Support for general targets (#8) (@ashryaagr)