v0.20.3

TMLE v0.20.3

Diff since v0.20.2

Merged pull requests:

• Prevalence weighted nuisance functions (#138) (@joshua-slaughter)
• Ol ccwtmle (#143) (@olivierlabayle)
• Update weighting procedure to handle nested model objects (#145) (@joshua-slaughter)

Closed issues:

• Case-Control Weighted TMLE (#136)

v0.20.2

TMLE v0.20.2

Diff since v0.20.1

v0.20.1

TMLE v0.20.1

Diff since v0.20.0

v0.20.0

TMLE v0.20.0

Diff since v0.19.0

Breaking changes

• Composition is now managed through an extension, a backend must be loaded together with DifferentiationInterface

Merged pull requests:

• JOSS Review Updates (#135) (@olivierlabayle)
• Fix manuscript and bibtex (#141) (@jbytecode)
• Ol dev doc (#142) (@olivierlabayle)

v0.19.0

TMLE v0.19.0

Diff since v0.18.1

Breaking changes

• Weighted TMLE is now the default
• Renaming: TMLEE -> Tmle, OSE -> Ose, NAIVE -> Plugin
• Datasets must be a DataFrame

New Features

• Multithreaded acceleration for nuisance function estimation
• Two C-TMLE flavours
• A new resampling strategy taking into account the treatment variables

Merged pull requests:

• Ol ctmle template (#125) (@olivierlabayle)
• Update README.md (#133) (@olivierlabayle)

Closed issues:

• Paper submission (#53)
• Multiple steps TMLE (#90)
• C-TMLE (#92)
• Documentation needs some love (#108)
• GLM for ill-formed matrices (#116)
• Bug if factorialEstimand is built from a dataset with categorical treatments (#124)
• Acceleration (#126)
• Greedy C-TMLE (#127)
• Documentation (#128)
• Cleaning (#129)
• Write JOSS paper and submit (#130)

v0.18.1

TMLE v0.18.1

Diff since v0.18.0

• Fix independence assumption in propensity score evaluation. Instead of assuming p(T1, ...Tp) = p(T1)*...*p(Tp), we now have a general factorization rule p(T1, ...Tp) = TT p(Ti|T<i).

Merged pull requests:

• Dependent ps (#123) (@olivierlabayle)

Closed issues:

• Link to documentation in repo description (#122)

v0.18.0

TMLE v0.18.0

Diff since v0.17.1

New Features

• Multiple steps TMLE

Breaking changes

• change from AbstractDifferentiation to DifferentiationInterface

Merged pull requests:

• Adds possibility to target more than once (#119) (@olivierlabayle)
• Changr from AD to DI (#121) (@olivierlabayle)

Closed issues:

• IPW and AIPW (#93)

v0.17.1

TMLE v0.17.1

Diff since v0.17.0

• New CausalTables @salbalkus
• Add documentation for YAML/JSON integrations @olivierlabayle

Merged pull requests:

• implement function to fit estimators on a CausalTable (#118) (@salbalkus)

Closed issues:

• Generate all parameters from a parameter family given data (#77)
• Find the time to rework docs (#95)
• Bring the FailedEstimate from TMLECI here (#113)

v0.17.0

TMLE v0.17.0

Diff since v0.16.1

Merged pull requests:

• Agnostic composed (#110) (@olivierlabayle)
• For 0.17.0 release (#112) (@olivierlabayle)

Closed issues:

• ComposedEstimands shouldn't take a function (#102)

v0.16.1

TMLE v0.16.1

Diff since v0.16.0

Merged pull requests:

• Cv check and doc (#109) (@olivierlabayle)