Changes in version 3.1.0 (2026-04-19)                  

Breaking changes

  - importance() is now a true S3 generic rather than an alias-style
    front end.
  - partial.ivarpro() has been replaced by plot.ivarpro().
  - The supported user-facing interfaces for fitted objects are now the
    corresponding S3 generics, such as importance(), predict(), and
    plot().

S3 interface cleanup

  - Registered importance() methods for "varpro" and "uvarpro" objects.
  - Registered plot() methods for "ivarpro" and "partialpro" objects.
  - Continued support for class-specific predict() methods through
    standard S3 dispatch for "varpro", "uvarpro", "ivarpro", and
    "isopro" objects.

Documentation

  - Help topics retain dotted method names such as plot.ivarpro,
    plot.partialpro, predict.ivarpro, predict.varpro, predict.uvarpro,
    and predict.isopro so that method pages remain easy to find in the
    reference manual and via ?topic.
  - Usage sections were updated to show S3 method signatures
    consistently, for example \method{plot}{ivarpro}(x, ...) and
    \method{predict}{ivarpro}(object, ...).
  - Examples were updated to use the generic forms plot(x, ...),
    predict(object, ...), and importance(object).
  - The iVarPro plotting documentation now uses data for the original
    feature matrix and documents target explicitly for multivariate and
    multiclass outputs.

Migration notes

  - Replace calls of the form partial.ivarpro(iv, var = ...) with
    plot(iv, var = ...).
  - Prefer importance(fit) over direct calls to importance.varpro(fit).
  - Prefer predict(fit, ...) over direct calls to predict.class(fit,
    ...).

                 Changes in version 3.0.0 (2026-04-08)                  

Improvements

  - Refactored varpro.strength() to reduce R-side post-processing
    overhead after the native varProStrength call, improving performance
    on large forests and large membership reconstructions.
  - Improved scalability and stability of varpro.strength(...,
    membership = TRUE) for very large analyses.
  - For RHF grow objects, varpro.strength() now uses the integrated
    hazard exposure values stored on the fitted object (int.haz.oob) as
    the default working response when available.
  - Internal cleanup of native-output decoding and membership
    reconstruction logic.

Bug fixes

  - Fixed a failure that could occur on very large analyses when
    rebuilding membership lists in R after native execution, which could
    previously surface as an integer-overflow warning from cumsum()
    followed by a downstream missing-value error in membership
    reconstruction.

                 Changes in version 2.1.0 (2026-02-12)                  

  - Major refactoring and enhancement to functions downstream from the
    entry varpro() function.

                 Changes in version 2.0.0 (2026-01-11)                  

  - Improved ivarPro.
  - Refactored code to improve speed.
  - Eliminated or replaced mclapply() with PSOCK-based parallel
    execution, improving Windows compatibility.

                 Changes in version 1.0.1 (2025-12-14)                  

  - CRAN compliance fixes.

                 Changes in version 1.0.0 (2025-12-11)                  

  - Initial release.