Skip to contents

hvtiBoostmtree 2.0.1

Bug Fixes

  • vimpPlot() no longer over-supplies bar labels. The internal graphics::text() calls passed rep(xvar.names, 3) — three labels for every bar — which triggered a 'labels' truncated to length ... warning and could mislabel bars. Labels now correspond one-to-one with the plotted bars.
  • plot.boostmtree() now draws the optimal-iteration reference line for every class of a multiclass (Nominal/Ordinal) prediction object. The internal abline() calls indexed an already-scalar Mopt as Mopt[q], yielding NA (and a silently missing line) for the second and later classes.

hvtiBoostmtree 2.0.0

Breaking Changes

  • Package renamed from boostmtree to hvtiBoostmtree (internal-only fork; the original boostmtree name remains with the upstream CRAN package). The boostmtree() function and S3 methods are unchanged — usage is now library(hvtiBoostmtree); boostmtree(...).
  • Minimum R version bumped to 4.1.0.
  • Maintainership transferred to John Ehrlinger. Original authors (Ishwaran, Pande, Kogalur) remain as authors.

New Features

  • Added ggRandomForests, hvtiPlotR, and hvtiRutilities as optional suggested dependencies for enhanced ggplot2-based visualization workflows.
  • Added two vignettes: Introduction to boostmtree and Longitudinal Data Analysis with boostmtree.
  • pkgdown documentation site fully restructured with reference index, articles, and sidebar badges.
  • BugReports field added to DESCRIPTION pointing to GitHub Issues.

Documentation

  • Comprehensive README rewritten with full badge set, installation instructions, quick-start examples, and reference citations.
  • NEWS.md introduced (in addition to inst/NEWS) for pkgdown changelog rendering.