Releases: jaredhuling/personalized
Releases · jaredhuling/personalized
CRAN release 0.2.7
- Bug fix for GAM-based losses
- Replaces gbm implementation with xgboost using a custom loss function
- Adds vignette for estimation of flexible ITRs via xgboost
CRAN Release 0.2.6
- Various small improvements/bug fixes
- Added changes to reflect the incoming JSS publication related to this package
- The DOI in the CITATION is for a new JSS publication that will be registered after
publication on CRAN.
CRAN Release 0.2.5
- Adds utilities for construction of augmentation functions and propensity score fitting functions
- Adds vignette for multi-category treatments and for usage of augmentation/propensity score utilities
- Various bug fixes
CRAN Release 0.2.4
- Added warning for use of Harrell's procedure in high dimensions
- Changed default value for 'train.fraction' to 0.75 from 0.5 in 'validate.subgroup'
- Minor improvements to plotting for 'subgroup_validated' objects
CRAN Release 0.2.3
- Fixes trt factor level reordering issue for plots
CRAN Release 0.2.2
- Fixes model printing error
- Improves subgroup.summarize() output
CRAN Release 0.2.1
- Fixes default argument bug for r-oldrel + windows
CRAN Release 0.2.0
- Simplified plot labeling
- Added clarifications to documentation
- Added customized loss function option
- Added options for count outcomes via Poisson negative log-likelihood as the loss
- Added treatment effect calculation based on estimated benefit scores
- Clarified/improved printing
- Improved numerical stability of weighted.ksvm
CRAN release 0.1.5
- Added printing of subgroup_validated results for quantile cutoffs via the which.quant argument
CRAN release 0.1.4
- add quantile and median cutpoints as options
- fixes subgroup effect calculation to account for weights
- add conditional plots for fitted and validated objects