Code

v0.9.0#143

Code

Merged

pzivich merged 62 commits intomasterpzivich/zEpid:masterfrom v0.9.0pzivich/zEpid:v0.9.0Copy head branch name to clipboardDec 30, 2020

+9,476-637

Commits

Commits on Jul 2, 2020

• adding functionality for DAG

  pzivich

  committed
  73376a1

  Copy full SHA for 73376a1

Commits on Jul 5, 2020

• adding direct input from networkx object

  pzivich

  committed
  db05662

  Copy full SHA for db05662

• adding functionality to load and optional networkx dependency

  pzivich

  committed
  7f64ac4

  Copy full SHA for 7f64ac4

• adding tests for causal graphs

  pzivich

  committed
  18bc89a

  Copy full SHA for 18bc89a

• adding better error handling

  pzivich

  committed
  3841f3e

  Copy full SHA for 3841f3e

• adding documentation for DirectedAcyclicGraph

  pzivich

  committed
  af43b80

  Copy full SHA for af43b80

• travis updates

  pzivich

  committed
  7b38350

  Copy full SHA for 7b38350

• versioning numpy on travis for python 3.5

  pzivich

  committed
  fc145d7

  Copy full SHA for fc145d7

Commits on Jul 12, 2020

• init 0.9.0 and new travis setup

  pzivich

  committed
  f2acde7

  Copy full SHA for f2acde7

• optional import in init

  pzivich

  committed
  d918231

  Copy full SHA for d918231

• Merge branch 'dag' into v0.9.0

  pzivich

  committed
  1e8635b

  Copy full SHA for 1e8635b

Commits on Jul 19, 2020

• Cross-fit AIPTW implementations

  pzivich

  committed
  deae6ee

  Copy full SHA for deae6ee

• adding tests for causal graphs

  pzivich

  committed
  a429d99

  Copy full SHA for a429d99

• aipw_calculator as function in utils

  pzivich

  committed
  ac548ef

  Copy full SHA for ac548ef

Commits on Aug 3, 2020

• fixing returning IDs via #126

  pzivich

  committed
  db7b8f5

  Copy full SHA for db7b8f5

• a working zipper plot implementation

  pzivich

  committed
  e715ba2

  Copy full SHA for e715ba2

Commits on Dec 12, 2020

• Adoption of SuperLearner into zEpid. Adds EmpiricalMeanSL and StepwiseSL as estimation functions following the sklearn syntax for use with SuperLearner

  pzivich

  committed
  900dce4

  Copy full SHA for 900dce4

• Finished both forward and backward step-wise selection for the models

  pzivich

  committed
  037eb1c

  Copy full SHA for 037eb1c

Commits on Dec 13, 2020

• SuperLearner implementation added. Still need to compare to R, write tests, and add docs to website

  pzivich

  committed
  1944271

  Copy full SHA for 1944271

• Added examples of use to documentation

  pzivich

  committed
  2c77db3

  Copy full SHA for 2c77db3

• Adding test to SuperLearner and the candidate estimators

  pzivich

  committed
  686dc31

  Copy full SHA for 686dc31

Commits on Dec 14, 2020

• Finished tests

  pzivich

  committed
  92f9a30

  Copy full SHA for 92f9a30

• Having SuperLearner return self so it works with the current setup of machine learning with TMLE and CrossFit

  pzivich

  committed
  f45da47

  Copy full SHA for f45da47

Commits on Dec 15, 2020

• adding custom_model to AIPW

  pzivich

  committed
  ef71e67

  Copy full SHA for ef71e67

• replacing bounding in TMLE with new format

  pzivich

  committed
  6569882

  Copy full SHA for 6569882

• further tweaks to superlearner

  pzivich

  committed
  4f3a1e6

  Copy full SHA for 4f3a1e6

• adding better tests for TMLE with custom_model

  pzivich

  committed
  744e1a9

  Copy full SHA for 744e1a9

• finishing replacement of _bounding_ with zepid.calc.utils.probability_bounds

  pzivich

  committed
  5d9e409

  Copy full SHA for 5d9e409

Commits on Dec 16, 2020

• updating readthedocs to have Super Learner documentation

  pzivich

  committed
  fc68556

  Copy full SHA for fc68556

• Adding risk-ratio support for AIPTW and CrossfitAIPTW

  pzivich

  committed
  fac0e07

  Copy full SHA for fac0e07

• Adding seeds for partitioning procedure for reproducibility

  pzivich

  committed
  297b9ba

  Copy full SHA for 297b9ba

• Adding example data set for the crossfit estimators

  pzivich

  committed
  8aee128

  Copy full SHA for 8aee128

• AIPTW cross-fitting fit tests. Just check that current operationalization is not broken by any future changes (do not have an outside or ground-truth comparison due to how randomness factors in)

  pzivich

  committed
  1b6f653

  Copy full SHA for 1b6f653

Commits on Dec 17, 2020

• adding CrossfitTMLE to cross-fitting estimators

  pzivich

  committed
  8fecc0f

  Copy full SHA for 8fecc0f

Commits on Dec 27, 2020

• fixing random_state flow

  pzivich

  committed
  69151dd

  Copy full SHA for 69151dd

• adding AIPTW OOB detector

  pzivich

  committed
  065f1a3

  Copy full SHA for 065f1a3

• adding additional superlearner probability bound check for predict_proba

  pzivich

  committed
  983428f

  Copy full SHA for 983428f

• adding diagnostics to cross-fit estimators

  pzivich

  committed
  a24ce59

  Copy full SHA for a24ce59

• updating setup.py to drop python 3.5 support

  pzivich

  committed
  386487c

  Copy full SHA for 386487c

• fixing AIPW warning to avoid misdetection for continuous outcomes

  pzivich

  committed
  8a7e7f8

  Copy full SHA for 8a7e7f8

• adding more tests for the cross-fit estimators

  pzivich

  committed
  7596e8a

  Copy full SHA for 7596e8a

• some test fixes. I think it will still fail due to seeds though

  pzivich

  committed
  f5ac31a

  Copy full SHA for f5ac31a

Commits on Dec 28, 2020

• should fix travisci issue...

  pzivich

  committed
  061f0b1

  Copy full SHA for 061f0b1

• Alternative way to specify RandomState for travisci compat?

  pzivich

  committed
  095aa13

  Copy full SHA for 095aa13

• checking if on numpy or pandas side for differences

  pzivich

  committed
  44a1e27

  Copy full SHA for 44a1e27

• well it isn't numpy, so maybe it is pandas...

  pzivich

  committed
  076458d

  Copy full SHA for 076458d

• the issue is reproducibility is not a function of the sample splitting by numpy/pandas, nor a result of version differences between numpy, pandas, scikit-learn

  pzivich

  committed
  984cf10

  Copy full SHA for 984cf10

• fix travis-requirements typo...

  pzivich

  committed
  a0bc459

  Copy full SHA for a0bc459

• adding GLM SuperLearner functionality (since sci-kit learn has extra stuff and doesn't always match statsmodels)

  pzivich

  committed
  cf9837b

  Copy full SHA for cf9837b

• Updating crossfit and other DR estimators to use GLMSL rather than sklearn

  pzivich

  committed
  c04ae71

  Copy full SHA for c04ae71

• fixing equal to all_close in GMLSL tests

  pzivich

  committed
  e819b47

  Copy full SHA for e819b47

Commits on Dec 29, 2020

• adding function to load formatted data (and checked it with IPTW)

  pzivich

  committed
  ae5cbee

  Copy full SHA for ae5cbee

• added binary exposure check to check_input_data as this is in GEstimationSNM

  pzivich

  committed
  839d267

  Copy full SHA for 839d267

• Replaced check_input_data in most estimators! (generalize will eventually need their own custom option)

  pzivich

  committed
  d5211de

  Copy full SHA for d5211de

• Better looking print_results returned

  pzivich

  committed
  51352b7

  Copy full SHA for 51352b7

• adding variance warning to IPTW confidence intervals for the ATT and the ATU

  pzivich

  committed
  948ea74

  Copy full SHA for 948ea74

• adding manual test for zipper_plot

  pzivich

  committed
  877bbbe

  Copy full SHA for 877bbbe

• adding zipper_plot to ReadTheDocs

  pzivich

  committed
  e3759b7

  Copy full SHA for e3759b7

• Adding GLMSL docs to ReadTheDocs

  pzivich

  committed
  5bb0dbc

  Copy full SHA for 5bb0dbc

Commits on Dec 30, 2020

• Updating ReadTheDocs (and making it easier to manage going forward)

  pzivich

  committed
  eccf829

  Copy full SHA for eccf829

• Update DAG ReadTheDocs

  pzivich

  committed
  07f46de

  Copy full SHA for 07f46de

• Updating Change-log for all v0.9.0 additions

  pzivich

  committed
  1451729

  Copy full SHA for 1451729