v1.6.1

• trajectories(lcModel) outputs cluster membership (#153)
• Improved visuals of trajectory plots (#152)
• Fixed dynamic variable handling (#151)

v1.6.0

latrend v1.6.0

• Resolved CRAN problems (#147)
• Re-added clusterCrit as dependency (#145)
• Ordering and labeling of cluster trajectories in plot (#144, #142)

v1.5.1

• Resolved CRAN errors and warnings
• Alternative implementation of ASW and Dunn index due to removal of clusterCrit package from CRAN (#131, #132)
• New PAP.adh dataset (#130)
• Added support to lcmm methods for (parallel) gridsearch (#126)
• More robust implementation of meta methods, allowing argument pass-through (#128)
• Improved package documentation (#127)

v1.5.0

New features

• Implemented virtual class for meta methods (#61)
• Implemented a meta method for repeatedly fitting a method and selecting the best fit (see lcFitRep, lcFitRepMin, lcFitRepMax) (#61)
• Implemented a meta method for repeatedly fitting a method until convergence (see lcFitConverged) (#61)

Major changes

• Made getLcMethod() generic

Other changes

• Added check for formula argument to LMKM method
• Added "converged" slot to lcModelPartition
• Generic validate() checks for correct output length
• Added workaround for erroneous R CMD check rmarkdown import note

v1.4.3

• Removed usage of soon-to-be deprecated ggplot functions

v1.4.2

Bug fixes

• Fixed minor bug in evaluate.lcMethod()

CRAN

• Resolved lcmm test error
• Shortened runtime of vignettes
• Shortened runtime of tests

v1.4.1

Bug fixes

• Fixed bug in predict.lcModel() introduced by #116
• Attempt to fix the documentation warning about how the escaped ampersand is not escaped

Breaking changes

• Default initialization of lcMethodLcmmGMM and lcMethodLcmmGBTM is set to init = "lme" to support lcmm v2.0.0 #121

v1.4.0

New features

• Added RMSE and WRMSE metrics
• plotMetric() now shows legend when grouping
• externalMetric(lcModels) returns a named distance matrix

Bug fixes

• Improved kml package compatibility; fixed errors on datasets with variable-length short trajectories
• Resolved warnings from latrendBoot() and latrendCV()

Breaking changes

• Removed redundant lcModelCustom in favor of lcModelPartition
• Renamed lcMethodCustom to lcMethodFunction for clarity

v1.3.1

• Added RMSE and WRMSE metrics (#106)
• Bugfix: latrendBoot and latrendCV drop empty trajectory identifier levels during sampling (#107)
• Implemented color scale and legend for grouped metric plot (#108)

v1.3.0

New features

1. Parameterized method testing framework for lcMethod and lcModel implementations. See test.latrend().
2. Warnings for missing data observations
3. Better handling of data with missing observations

Breaking changes

1. Removed support for the longclust package as it is no longer available on CRAN.

Other changes

1. Updated examples, vignettes and tests to pass with _R_CHECK_DEPENDS_ONLY_ = true.