Releases 0.3.0
Use active bindings in the R6 OOP implementation #106 & #93
Fix the aggregation formula for standard errors from repeated cross-fitting #94 & #95
Always use the same bootstrap algorithm independent of dml1 vs dml2 and consistent with docu and paper #98 & #99
Initialize predictions with NA and make sure that there are no missleading entries in the evaluated score functions #96 & #105
Avoid overriding learner parameters during turing #83 & #84
Fixes in the exception handling and extension of the unit tests for the score function choice #82
Prevent overwriting parameters from initialization when calling set_ml_nuisance_params #87 & #89
Major refactoring and cleanup and extension of the unit test framework #101
Extension and reorganization of exception handling for DoubleMLData objects #63 & #90
Introduce style guide and clean up code #80 & #81
Adaption to be compatible with an API change in the next mlr3 release #103
Run unit tests with mlr3 in dev version on github actions #104
Updated the citation info #78 , #79 & #86
Added a short version of and a reference to the arXiv paper as vignette #110 & #113
Prevent using the subclassed methods check_score and check_data when constructing DoubleML objects #107
Other refactoring and minor adaptions #91 , #92 , #102 & #108
You can’t perform that action at this time.
