Models can be improved by post-processing class probabilities, by: recalibration, conversion to hard probabilities, assessment of equivocal zones, and other activities. 'probably' contains tools for conducting these operations.
| Version: | 0.0.2 |
| Depends: | R (≥ 2.10) |
| Imports: | dplyr, generics, rlang, tidyselect, vctrs, yardstick |
| Suggests: | covr, ggplot2, knitr, parsnip, rmarkdown, rsample, testthat |
| Published: | 2019-03-07 |
| Author: | Max Kuhn [aut, cre], Davis Vaughan [aut], RStudio [cph] |
| Maintainer: | Max Kuhn <max at rstudio.com> |
| BugReports: | https://github.com/tidymodels/probably/issues |
| License: | GPL-2 |
| URL: | https://github.com/tidymodels/probably/ |
| NeedsCompilation: | no |
| Materials: | README NEWS |
| CRAN checks: | probably results |
| Reference manual: | probably.pdf |
| Vignettes: |
Equivocal zones Where does probably fit in? |
| Package source: | probably_0.0.2.tar.gz |
| Windows binaries: | r-devel: probably_0.0.2.zip, r-release: probably_0.0.2.zip, r-oldrel: probably_0.0.2.zip |
| OS X binaries: | r-release: probably_0.0.2.tgz, r-oldrel: probably_0.0.2.tgz |
| Old sources: | probably archive |
Please use the canonical form https://CRAN.R-project.org/package=probably to link to this page.