Skip to content

Releases: ing-bank/probatus

probatus 1.7.0

16 Mar 09:59
Compare
Choose a tag to compare

Improvements in this release:

  • Create a comparison of imputation strategies #86
  • Added support for passing check_additivity argument #103
  • Range of code styling issues fixed, based on precommit config #100
  • Renamed TreeDependencePlotter to DependencePlotter and exposed the docs #94
  • Enable instalation of extra dependencies #97
  • Added how to notebook to ensure reproducibility #99
  • Description of vision of probatus #91

probatus 1.6.2

10 Mar 11:45
Compare
Choose a tag to compare

Improvements in this release:

  • Bugfix, allow passing kwargs to dependence plot in ShapModelInterpreter #90

probatus 1.6.1

09 Mar 12:51
1595cab
Compare
Choose a tag to compare

Improvements in this release:

  • ShapRFECV supports all sklearn compatible search CVs. #76 #49

probatus 1.6.0

01 Mar 12:34
Compare
Choose a tag to compare

Improvements in this release:

  • Added features list to README #53
  • Added docs for sample row functionality #54
  • Added 'open in colab' badges to tutorial notebooks #56
  • Deploy documentation on release #47
  • Added columns_to_keep for shap feature elimination #63
  • Updated docs for usage of columns to keep functionality in SHAPRFECV #66
  • Added shap support for linear models #69
  • Installed probatus in colab notebooks #80
  • Minor infrastructure tweaks #81

probatus 1.5.1

04 Dec 12:03
Compare
Choose a tag to compare

Various improvements to the consistency and usability of the package

  • Unit test docstring and notebooks #41
  • Unified scoring metric within probatus #27
  • Improve docstrings consistency documentation #25
  • Implemented unified interface #24
  • Added images to API docs documentation #23
  • Added verbose parameter to ShapRFECV #21
  • Make API more consistent #19
    • Set model parameter name to clf across probatus
    • Set default random_state to None
    • Ensure that verbose is used consistently in probatus
    • Unify parameter class_names for classes in which it is relevant
    • Add return scores parameter to compute wherever applicable
  • Add sample row functionality to utils #17
  • Make an experiment comparing sklearn.RFECV with ShapRFECV #16
  • ShapModelInterpreter calculate train set feature importance #13

probatus 1.5.0

18 Nov 10:13
Compare
Choose a tag to compare

Improved

  • Docs and API of ShapRFECV #11