{msm.stacked} 0.0.3

This release of {msm.stacked} included changes that are described below.

Breaking changes

• Updated algorithm to estimate confidence intervals for transition probabilities to be more consistent across time-points.
  This is based on a parametric bootstrap approach, where we use each resample along the entire time requested for prediction, with confidence intervals then calculated using the percentile method.
  The arguments of stacked.data.msm() have been updated accordingly, and names have been adjusted to not collide with {msm}.
  Please read the documentation page at ?stacked.data.msm() to get familiar with the new arguments.

Housekeeping

• Removed a URL in the vignette on confidence intervals that was giving a 403 Forbidden error.

{msm.stacked} 0.0.2

This releases of {msm.stacked} includes changes that are described below.

New features

• It is now possible to calculate confidence intervals for the transition probabilities returned by stacked.data.msm() via the ci argument.
  This supports all methods implemented in msm::pmatrix.msm(), and can be used with multi-state models with and without covariates.
• Added new vignettes describing the functionality of {msm.stacked}.
  They can be found by typing the following in your R console:

  vignette("A-introduction", package = "msm.stacked")
  vignette("B-ci", package = "msm.stacked")

Housekeeping

• Updated README file.
• Updated copyright year.

{msm.stacked} 0.0.1

Initial release of the package. Currently, the following functions are included:

• stacked.data.msm(), to calculate transition probabilities over time from an {msm} model fit;
• stacked.plot.msm(), to automatically produce stacked probabilities plots based on transition probabilities over time;
• states.msm(), a utility function to determine the names of states for an {msm} model.