`densityMclust()`

by default draw a graph of the density estimate.- Fixed a bug in computing mixture density if the noise component is present.
- Changed default behaviour of
`hc()`

when called to perform agglomerative hierarchical clustering instead of using for EM initialization. - The default
`mclust.options("hcModelName")`

now returns only the model to be used. - Changed default
`partition`

argument of`hc()`

function by adding`dupPartion()`

to remove data duplicates. - Added checks to
`mclustBootstrapLRT()`

to stop if an invalid`modelName`

is provided or a one-component mixture model is provided. - Extended the functionality of
`cvMclustDA()`

by including as cross-validated metrics both the classification error and the Brier score. - Updated info on dataset man pages.

- Updated plot method (dendrogram) for hierarchical clustering — now based on classification likelihood.
- Added
`MclustSSC()`

function (and related`print`

,`summary`

,`plot`

, and`predict`

, methods) for semi-supervised classification. - Exchanged order of models VEE and EVE to account for increasing complexity of EVE.
- Added
`cex`

argument to`clPairs()`

to control character expansion used in plotting symbols. `em()`

and`me()`

have now`data`

as first argument.

- Fixed issues with source Fortran code with gfortran 10 as reported by CRAN.
- Clean code of
`hcCriterion()`

. - Replaced
`CEX`

argument in functions with standard base graph`cex`

argument. - Removed
`ylim`

argument in function; it can be passed via`...`

. - MclustDA models use the default SVD transformation of the data for initialisation of the EM algorithm.
- Added
`icl`

criterion to object returned by`Mclust()`

. - Fixed number of pages for the RJ reference.
- quantileMclust() uses bisection line search method for numerically computing quantiles.

- Fixed warnings in Fortran calls raised by CRAN.

- Added
`classPriorProbs()`

to estimate prior class probabilities. - Added
`BrierScore()`

to compute the Brier score for assessing the accuracy of probabilistic predictions. - Added
`randomOrthogonalMatrix()`

to generate random orthogonal basis matrices. - Partial rewriting of
`summary.MclustDA()`

internals to provide both the classification error and the Brier score for training and/or test data. - Partial rewriting of
`plot.MclustDA()`

internals. - Added
`dmvnorm()`

for computing the density of a general multivariate Gaussian distribution via efficient Fortran code. - Added Wisconsin diagnostic breast cancer (WDBC) data.
- Added EuroUnemployment data.
- Fixed mismatches in Fortran calls.
- Bugs fix.

- Added website site and update DESCRIPTION with URL.
- Fixed a bug when checking for univariate data with a single observation in several instances. Using
`NCOL()`

works both for n-values vector or nx1 matrix. - Fixed a bug when
`hcPairs`

are provided in the`initialization`

argument of`mclustBIC()`

(and relatives) and the number of observations exceed the threshold for subsetting. - Fixed bugs on axes for some manual pairs plots.
- Renamed
`type = "level"`

to`type = "hdr"`

, and`level.prob`

to`prob`

, in`surfacePlot()`

for getting HDRs graphs - Fixed a bug in
`type = "hdr"`

plot on`surfacePlot()`

. - Fixed a bug in
`as.Mclust()`

. - Small changes to
`summary.MclustDA()`

when`modelType = "EDDA"`

and in general for a more compact output.

- Added
`mclustBICupdate()`

to merge the best values from two BIC results as returned by`mclustBIC()`

. - Added
`mclustLoglik()`

to compute the maximal log-likelihood values from BIC results as returned by`mclustBIC()`

. - Added option
`type = "level"`

to`plot.densityMclust()`

and`surfacePlot()`

to draw highest density regions. - Added
`meXXI()`

and`meXXX()`

to exported functions. - Updated vignette.

- Added parametric bootstrap option (
`type = "pb"`

) in`MclustBootstrap()`

. - Added the options to get averages of resampling distributions in
`summary.MclustBootstrap()`

and to plot resampling-based confidence intervals in`plot.MclustBootstrap()`

. - Added function
`catwrap()`

for wrapping printed lines at`getOption("width")`

when using`cat()`

. `mclust.options()`

now modify the variable`.mclust`

in the namespace of the package, so it should work even inside an mclust-function call.- Fixed a bug in
`covw()`

when`normalize = TRUE`

. - Fixed a bug in
`estepVEV()`

and`estepVEE()`

when parameters contains`Vinv`

. - Fixed a bug in
`plotDensityMclustd()`

when drawing marginal axes. - Fixed a bug in
`summary.MclustDA()`

when computing classification error in the extreme case of a minor class of assignment. - Fixed a bug in the initialisation of
`mclustBIC()`

when a noise component is present for 1-dimensional data. - Fixed bugs in some examples documenting
`clustCombi()`

and related functions.

- Model-based hierarchical clustering used to start the EM-algorithm is now based on the scaled SVD transformation proposed by Scrucca and Raftery (2016). This change is not backward compatible. However, previous results can be easily obtained by issuing the command:
`mclust.options(hcUse = "VARS") For more details see help("mclust.options")`

. - Added
`subset`

parameter in`mclust.options()`

to control the maximal sample size to be used in the initial model-based hierarchical phase. `predict.densityMclust()`

can optionally returns the density on a logarithm scale.- Removed normalization of mixing proportions for new models in single mstep.
- Internal rewrite of code used by
`packageStartupMessage()`

. - Fixed a small bug in
`MclustBootstrap()`

in the univariate data case. - Fixed bugs when both the noise and subset are provided for initialization.
- Vignette updated to include references, startup message, css style, etc.
- Various bug fixes in plotting methods when noise is present.
- Updated references in
`citation()`

and man pages.

- Added
`gmmhd()`

function and relative methods. - Added
`MclustDRsubsel()`

function and relative methods. - Added option to use subset in the hierarchical initialization step when a noise component is present.
`plot.clustCombi()`

presents a menu in interactive sessions, no more need of data for classification plots but extract the data from the`clustCombi`

object.- Added
`combiTree()`

plot for`clustCombi`

objects. `clPairs()`

now produces a single scatterplot in the bivariate case.- Fixed a bug in
`imputeData()`

when seed is provided. Now if a seed is provided the data matrix is reproducible. - in
`imputeData()`

and`imputePairs()`

some name of arguments have been modified to be coherent with the rest of the package. - Added functions
`matchCluster()`

and`majorityVote()`

. - Rewrite of print and summary methods for
`clustCombi`

class objects. - Added
`clustCombiOptim()`

. - Fixed a bug in
`randomPairs()`

when nrow of input data is odd. - Fixed a bug in
`plotDensityMclust2()`

,`plotDensityMclustd()`

and`surfacePlot()`

when a noise component is present.

- Added native routine registration for Fortran code.
- Fixed lowercase argument PACKAGE in
`.Fortran()`

calls.

- Fixed a bug in rare case when performing an extra M step at the end of EM algorithm.

- Replaced
`structure(NULL, *)`

with`structure(list(), *)`

- Added argument
`x`

to`Mclust()`

to use BIC values from previous computations to avoid recomputing for the same models. The same argument and functionality was already available in`mclustBIC()`

. - Added argument
`x`

to`mclustICL()`

to use ICL values from previous computations to avoid recomputing for the same models. - Fixed a bug on
`plot.MclustBootstrap()`

for the`"mean"`

and`"var"`

in the univariate case. - Fixed uncertainty plots.
- Added functions
`as.Mclust()`

and`as.densityMclust()`

to convert object to specific mclust classes. - Solved a numerical accuracy problem in
`qclass()`

when the scale of`x`

is (very) large by making the tolerance eps scale dependent. - Use transpose subroutine instead of non-Fortran 77 TRANSPOSE function in
`mclustaddson.f`

. - Fixed
`predict.Mclust()`

and`predict.MclustDR()`

by implementing a more efficient and accurate algorithm for computing the densities.

- Fixed slow convergence for VVE and EVE models.
- Fixed a bug in orientation for model VEE.
- Added an extra M-step and parameters update in
`Mclust()`

call via`summaryMclustBIC()`

.

- Added option to
`MclustBootstrap()`

for using weighted likelihood bootstrap. - Added a plot method for
`MclustBootstrap`

objects. - Added
`errorBars()`

function. - Added
`clPairsLegend()`

function. - Added
`covw()`

function. - Fixed rescaling of mixing probabilities in new models.
- bug fixes.

- Fixed bugs.
- Added print method for
`hc`

objects.

- Added the four missing models (EVV, VEE, EVE, VVE) to the mclust family. A noise component is allowed, but no prior is available.
- Added
`mclustBootstrapLRT()`

function (and corresponding print and plot methods) for selecting the number of mixture components based on the sequential bootstrap likelihood ratio test. - Added
`MclustBootstrap()`

function (and corresponding print and summary methods) for performing bootstrap inference. This provides standard errors for parameters and confidence intervals. - Added
`"A quick tour of mclust"`

vignette as html generated using rmarkdown and knitr. Older vignettes are included as other documentation for the package. - Modified arguments to
`mvn2plot()`

to control colour, lty, lwd, and pch of ellipses and mean point. - Added functions
`emX()`

,`emXII()`

,`emXXI()`

,`emXXX()`

,`cdensX()`

,`cdensXII()`

,`cdensXXI()`

, and`cdensXXX()`

, to deal with single-component cases, so calling the em function works even if`G = 1`

. - Small changes to
`icl()`

, now it is a generic method, with specialized methods for`Mclust`

and`MclustDA`

objects. - Fixed bug for transformations in the initialization step when some variables are constant (i.e. the variance is zero) or a one-dimensional data is provided.
- Changed the order of arguments in
`hc()`

(and all the functions calling it). - Small modification to
`CITATION`

file upon request of CRAN maintainers. - Various bug fixes.

- Added option for using transformation of variables in the hierarchical initialization step.
- Added
`quantileMclust()`

for computing the quantiles from a univariate Gaussian mixture distribution. - Fixed bugs on
`summaryMclustBIC()`

,`summaryMclustBICn()`

,`Mclust()`

to return a matrix of 1s on a single column for`z`

even in the case of`G = 1`

. This is to avoid error on some plots. - Moved pdf files (previously included as vignettes) to
`inst/doc`

with corresponding`index.html`

.

- Fixed bug for
`logLik.MclustDA()`

in the univariate case. - Added argument
`"what"`

to`predict.densityMclust()`

function for choosing what to retrieve, the mixture density or component density. `hc()`

function has an additional parameter to control if the original variables or a transformation of them should be used for hierarchical clustering.- Added
`"hcUse"`

argument in`mclust.options()`

to be passed as default to`hc()`

. - Added the storing of original data (and class for classification models) in the object returned by the main functions.
- Added component
`hypvol`

to`Mclust`

object which provide the hypervolume of the noise component when required, otherwise is set to NA. - Added a warning when prior is used and BIC returns NAs.
- Fixed bugs in
`summary.Mclust()`

,`print.summary.Mclust()`

,`plot.Mclust()`

and`icl()`

in the case of presence of a noise component. - Fixed bug on some plots in
`plot.MclustDR()`

which requires`plot.new()`

before calling`plot.window()`

. - Fixed a bug in
`MclustDR()`

for the one-dimensional case. - Corrections to
`Mclust`

man page. - Various small bug fixes.

- Fixed bug in
`sim*()`

functions when no obs are assigned to a component. `MclustDA()`

allows to fit a single class model.- Fixex bug in
`summary.Mclust()`

when a subset is used for initialization. - Fixed a bug in the function
`qclass()`

when ties are present in quantiles, so it always return the required number of classes. - Various small bug fixes.

- Added
`icl()`

function for computing the integrated complete-data likelihood. - Added
`mclustICL()`

function with associated print and plot methods. `print.mclustBIC()`

shows also the top models based on BIC.- Modified
`summary.Mclust()`

to return also the icl. - Rewrite of
`adjustedRandIndex()`

function. This version is more efficient for large vectors. - Updated help for
`adjustedRandIndex()`

. - Modifications to
`MclustDR()`

and its summary method. - Changed behavior of
`plot.MclustDR(..., what = "contour")`

. - Improved plot of uncertainty for
`plot.MclustDR(..., what = "boundaries")`

. - Corrected a bug for malformed GvHD data.
- Corrected version of qclass for selecting initial values in case of 1D data when successive quantiles coincide.
- Corrected version of plot BIC values when only a single G component models are fitted.
- Various bug fixes.

- Added new summary and print methods for
`Mclust()`

. - Added new summary and print methods for
`densityMclust()`

. - Included
`MclustDA()`

function and methods. - Included
`MclustDR()`

function and methods. - Included
`me.weighted()`

function. - Restored hierarchical clustering capability for the EEE model (hcEEE).
- Included vignettes for mclust version 4 from Technical Report No. 597 and for using weights in mclust.
- Adoption of GPL (>= 2) license.

- Added
`summary.Mclust()`

. - New functions for plotting and summarizing density estimation.
- Various bug fixes.
- Added
`clustCombi()`

and related functions (code and doc provided by Jean-Patrick Baudry). - Bug fix: variable names lost when G = 1.

- Added
`NAMESPACE`

.

- Removed intrinsic gamma-

- Fixed
`hypvol()`

function to avoid overflow. - Fixed
`hypvol()`

help file value description. - Removed unused variables and tabs from source code.
- Switched to intrinsic gamma in source code.
- Fixed default warning in estepVEV and mstepVEV.

- Fixed output when G = 1 (it had NA for the missing
`z`

component).

- Removed hierarchical clustering capability for the
`EEE`

model (hcEEE). - The R 2.12.0 build failed due to a 32-bit Windows compiler error, forcing removal of the underlying Fortran code for hcEEE from the package, which does not contain errors and compiles on other platforms.

- Added description of parameters output component to
`Mclust`

and`summary.mclustBIC`

help files.

- Added
`densityMclust()`

function.

- Fixed bug in covariance matrix output for EEV and VEV models.

- Bug fixes.

- Moved CITATION to inst and used standard format
- BibTex entries are in inst/cite.
- Fixed bug in handling missing classes in
`mclustBIC()`

. - Clarified license wording.

- Corrected output description in
`mclustModel`

help file. - Updated mclust manual reference to show revision.

- Updated
`defaultPrior`

help file. - Added utility functions for imputing missing data with the mix package.
- Changed default max to number of mixture components in each class from 9 to 3.

- Fixed problems with in
`mclustOptions`

help file

- Fixed
`plot.mclustBIC()`

and`plot.Mclust()`

to handle`modelNames`

. - Changed “orientation” for VEV, VVV models to be consistent with R
`eigen()`

and the literature - Fixed some problems including doc for the noise option.
- Updated the
`unmap()`

function to optionally include missing groups.

- Fixed bug in the
`"errors"`

option for`randProj()`

. - Fixed boundary cases for the
`"noise"`

option.

- Added permission for CRAN distribution to LICENSE.
- Fixed problems with help files found by new parser.
- Changed PKG_LIBS order in src/Makevars.
- Fixed
`Mclust()`

to handle sampling in data expression in call.

- Added
`EXPR = to`

all switch functions that didn’t already have it.

- Added
`pro`

component to parameters in`dens()`

help file. - Fixed some problems with the noise option.

- Default seed changed in
`sim*()`

functions. - Added model name check to various functions.
- Otherwise backward compatible with version 3.0

- Most plotting functions changed to use color.
`Mclust()`

and`mclustBIC()`

fixed to work with G=1- Otherwise backward compatible with version 3.0.

- New functionality added, including conjugate priors for Bayesian regularization.
- Backward compatibility is not guaranteed since the implementation of some functions has changed to make them easier to use or maintain.