Skip to content

Latest commit

 

History

History
45 lines (33 loc) · 1.68 KB

NEWS.md

File metadata and controls

45 lines (33 loc) · 1.68 KB

uniCATE 0.4.0 (2022-02-07)

  • When the super learning arguments are set to NULL in unicate() and sunicate(), cross-validated elastic net regressions are fit to estimate nuisance parameters instead of the previously used default sl3 super learners. This speeds up computation time, and makes it easier to apply these methods to small datasets by default. This doesn't affect the asymptotic behaviour of the unicate() estimator, either.

uniCATE 0.3.0 (2022-02-01)

  • uniCATE is going public!

uniCATE 0.2.3 (2022-01-31)

  • Adding GitHub Actions for CI and CD.
  • Made largely stylistic updates to internal functions to pass linting and style checks.
  • Set the license to Apache 2.0

uniCATE 0.2.2 (2022-01-31)

  • failure argument in sunicate() has been changed to event.
  • Change to sunicate() behaviour: The median of the data argument's relative_time variable is now set as the default time_cutoff when time_cutoff is NULL.
  • Change to internal sunicate() behaviour: When transforming wide data to a long format, no more than five unique relative times are reported for each observation. These relative times correspond to the quintiles of the relative times between the earliest relative time, and the minimum of each observations relative time and the time cutoff.

uniCATE 0.2.1 (2022-01-21)

  • Documentation of internal and exported functions has been cleaned.
  • New package introduction in README.
  • Co-authors added to DESCRIPTION
  • Package description updated in DESCRIPTION

uniCATE 0.2.0 (2022-01-18)

  • Added sunicate(), the unicate() counterpart for right-censored time-to-event outcomes.
  • Minor touch-ups to miscellaneous docs.