v0.8.1

[pzivich]

v0.8.1

TimeFixedGFormula

• [x] Allow estimation of ATT and ATU #110

• [x] Adding option to fit to either predict censored outcome or ignore them

• [x] Drop all missing data besides missing outcome data in init. This will prevent later pandas errors that are non-informative errors

IPTW

• [ ] Allows estimation of continuous treatments in IPTW #67

• Consider adding SurvivalIPTW in this update (add lifelines as a dependency?)

• Consider adding LongitudinalIPTW

AIPTW

• Consider adding SurvivalAIPTW

IPSW and AIPSW

• [x] Add weights for confounding into the class objects themselves

• [ ] Add weights for censoring into the class objects themselves

Smart Q-models #119

• [x] Catch missing treatment exposure in TimeFixedGFormula.outcome_model

• [x] Catch missing treatment exposure in SurvivalGFormula.outcome_model

• [x] Catch missing treatment exposure in IPTW.marginal_structural_model

• [x] Catch missing treatment exposure in GEstimationSNM.structural_nested_model

• [x] Catch missing treatment exposure in TMLE.outcome_model

• [x] Catch missing treatment exposure in AIPTW.outcome_model

• [x] Catch missing treatment exposure in GTransportFormula.outcome_model

• [x] Catch missing treatment exposure in AIPSW.outcome_model

[pzivich]

Since I am adding the IPTW estimation functionality to IPSW, I am going to move the weight calculation for IPTW to a utils.py

This change will mostly be behind-the-scenes re-structuring