drtmle 1.1.1

• Version 1.1.1 released on GitHub and CRAN.
• Major changes
  • fixed bug in how multi-level treatment estimates are computed
• Minor changes
  • fixed small bugs that caused errors predicting from some super learner wrapper functions.
  • removed earth dependency to fix build errors

drtmle 1.1.0

• Version 1.1.0 released on GitHub and CRAN.
• Major changes
  • add support for averaging over multiple super learner runs
  • add support for partially/fully cross-validated standard error estimates
  • default for stratify argument changed to FALSE
  • major additions to vignette documenting new features
• Minor changes
  • future parallel backend must be specified externally by user
  • accordingly, some dependencies have been updated

drtmle 1.0.5

As of January 2019: * Version 1.0.5 released on GitHub and CRAN. * Addition of snow and data.table packages to suggests to quell re-occurring warnings and errors on CRAN builds. * Added nSL option for fitting and averaging multiple super learners as part of the estimation procedure. * Added adapt_g option for outcome-adaptive propensity score fitting.

drtmle 1.0.4

As of August 2019: * Added minor touch-ups and link fixes to documentation and vignettes. * Improved how slots are ordered upon being included in the return object.

As of July 2019: * Version 1.0.4.9001 released on GitHub. * Removed dependency on plyr package. * Added option to average over repeated Super Learner fits.

As of December 18, 2018: * Version 1.0.4 released on GitHub and CRAN. * Resolved issues arising from returnModels option when users input nuisance parameters. * Added option to bypass future parallelization calls for easier debugging. * Fixed bugs in standard TMLE implementation – namely, more robust fluctuations and corrected variance estimators.

drtmle 1.0.3

As of July 2, 2018: * Version 1.0.3 released on GitHub and CRAN. * Fixed warnings on CRAN builds.

drtmle 1.0.2

As of February 5, 2018: * Version 1.0.2 released on GitHub and CRAN. * Replaced foreach parallelization with future. * Included more robust Super Learner methods. * Fixed test to pass build with long doubles removed. * Accommodated returning estimated influence functions with drtmle() fit for power users. * Incorporated minor documentation corrections and updates.

As of December 11, 2017: * Version 1.0.2.9000 released on GitHub. * More robust convex combination SuperLearner implemented.

drtmle 1.0.0

As of August 17, 2017: * Version 1.0.0 released on CRAN. * Version 1.0.0.9000 released on GitHub.

As of August 15, 2017: * Version 1.0.0 ready for CRAN release.

drtmle 0.0.1

As of April 05, 2017: * The first public release of this package (v. 0.0.1) is made available on GitHub.