v0.9.6

Release v0.9.6
https://pypi.org/project/causallib/0.9.6/

What's Changed

• Add **kwargs to covariate_imbalance_count_error by @mmdanziger in #64
• Adjust for pandas>2 and networkx>3 dependencies by @ehudkr in #65
• Test on Python 3.10 and 3.11 by @ehudkr in #65

Full Changelog: v0.9.5...v0.9.6

v0.9.5

Release v0.9.5
https://pypi.org/project/causallib/0.9.5/

What's Changed

• Count number of Imbalanced covariates metrics by @ehudkr in #59
• sklearn scorer wrapper for propensity models by @ehudkr in #59
• Align signature of causal metrics with **kwargs by @ehudkr in #59
• Add a name to the time variable Series in NHEFS survival data by @ehudkr in #59

Full Changelog: v0.9.4...v0.9.5

v0.9.4

Release v0.9.4
https://pypi.org/project/causallib/0.9.4/

What's Changed

• NHEFS survival data now respects augmentation parameters by @ehudkr in #57
• Allow different covariate specification to weight and outcome models in WeightedStandardizedSurvival by @ehudkr in #57
• Allow specifying no covariates (e.g., intercept-only models) in doubly robust (and WeightedStandardizedSurvival) models by @ehudkr in #57
• Drop the numpy <1.24.0 dependency restriction by @ehudkr in #57

Full Changelog: v0.9.3...v0.9.4

v0.9.3

Release v0.9.3
https://pypi.org/project/causallib/0.9.3/

What's Changed

• Fix individual survival curves estimation bug by @ehudkr in #56

Full Changelog: v0.9.2...v0.9.3

v0.9.2

Release v0.9.2
https://pypi.org/project/causallib/0.9.2/

What's Changed

• Constrain Scikit-learn dependency <1.2 until a solution @ehudkr in #51
• Align Matching to WeightEstimator interface by @ehudkr in #53
• Optional direct labeling in covariate balance plots by @ehudkr in #54

Full Changelog: v0.9.1...v0.9.2

v0.9.1

Release v0.9.1
https://pypi.org/project/causallib/0.9.1/

What's Changed

• Model selection within weight-based survival models by @ehudkr in #47

Full Changelog: v0.9.0...v0.9.1

v0.9.0

Release v0.9.0
https://pypi.org/project/causallib/0.9.0/

Main changes

Two main additions on the model evaluations front.

1. We refactored the whole evaluation module, changing the API to be a lot more user friendly, with options to customize the generated plots.
2. We added a whole suite of causal-oriented metrics and scorers, that allow to integrate with scikit-learn's model selection machinery (like GridSearchCV, or any other scikit-learn compatible hyperparameter search model), and perform model selection in cross validation.

What's Changed

• limit n_neighbors to n_samples before matching by @mmdanziger in #38
• Evaluation refactoring and interface change by @mmdanziger in #40
• Covariate imbalance scatterplot by @edenjenzohar in #43
• Causal model selection by @ehudkr in #45

New Contributors

• @edenjenzohar made their first contribution in #43

Full Changelog: v0.8.2...v0.9.0

v0.8.2

Release v0.8.2
https://pypi.org/project/causallib/0.8.2/

What's Changed

• PropensityFeatureStandardization deepcopy fix by @mmdanziger in #35

Full Changelog: v0.8.1...v0.8.2

v0.8.1

Release v0.8.1
https://pypi.org/project/causallib/0.8.1/

What's Changed

• Fix argument misalignment when passing custom metric to OutcomeEvaluator by @yoavkt in #33

Full Changelog: v0.8.0...v0.8.1

v0.8.0

Release v0.8.0
https://pypi.org/project/causallib/0.8.0/

What's Added:

• Causal survival models by @liorness in #25
• Confounder selection module by @ehudkr and @onkarbhardwaj in #22
• Targeted Maximum Likelihood Estimator (TMLE) by @ehudkr in #26
• Augmented Inverse Probability Weighting (AIPW) by @ehudkr in #30
• Multiple types of propensity-based features in doubly robust models by @ehudkr in #28 and #30
• R-learner by @Itaymanes in #24
• X-learner by @yoavkt in #31
• Verbosity control in IPW truncation by @liranszlak in #27

Backward compatibility-breaking changes

• Doubly robust models have been renamed @ehudkr in #28 and #30
  • DoublyRobustIpFeature to PropensityFeatureStandardization
  • DoublyRobustJoffe to WeightedStandardization
  • DoublyRobustVanilla to AIPW
• Asymmetric propensity truncation in IPW by @liranszlak in #27
  • Moving from a single symmetric truncation (truncate_eps) to a two-parameter asymmetric truncation (clip_min, clip_max)

New Contributors

• @onkarbhardwaj made their first contribution in #22
• @Itaymanes made their first contribution in #24
• @liorness made their first contribution in #25
• @liranszlak made their first contribution in #27
• @yoavkt made their first contribution in #31

Full Changelog: v0.7.1...v0.8.0