Changes in version 2.6.0 o imputeBDLs() has been renamed to imputeBDL(). imputeBDLs() is kept as a deprecated alias and will be removed in a future release. o impRZilr() is deprecated and is now an alias of imputeBDL() (the same method: EM-based replacement of rounded zeros in pivot coordinates). Its results are unchanged; please switch to imputeBDL(). o imputeBDL() now correctly imputes values below the detection limit even when they are supplied as small positive numbers rather than exact zeros. Earlier imputeBDLs() silently skipped such values because the checkData() thresholding result was discarded. Results for the documented usage (rounded zeros pre-coded as 0) are unchanged. o impAll() no longer returns NULL for data containing rounded zeros (it read a list element that no longer exists); it now calls imputeBDL() internally. o removed a stray diagnostic print() from the rounded-zero imputation, a duplicate internal impAll() definition, and broken non-default code paths (the 'test' argument of imputeBDLs() and the unused 'variation' branch of impRZilr()). o imputeUDLs() now returns an object of class "replaced", and the imputed-zero index ('wind') of imputeBDL()/imputeUDLs() is returned aligned with the original column order. Changes in version 2.5.0 o new function cellPcaCoDa() for cellwise-robust PCA of compositional data, with print/summary/plot methods, implementing Templ (2026, Advances in Data Analysis and Classification) o new function contaminate_simplex() to generate cellwise-contaminated compositions o cellPcaCoDa() defaults now follow the paper: Tukey biweight loss and MCD initialisation (new 'init' argument); the cellwise detection step uses the standardised projection test. Default-path results therefore differ from earlier development versions. o cellPcaCoDa() now requires at least 3 parts (D >= 3) and warns when k = D-1, where the residual space is empty and cellwise detection is disabled (use k < D-1). o cellPcaCoDa() detects period-2 limit cycles in the alternating scheme and stops at the stabilised cycle (new 'cycled' flag in the result) instead of running to maxiter. o ggplot2, pls and data.table moved from 'Depends' to 'Imports' (they are no longer attached to the search path; load them explicitly if your own code relied on that attachment) Changes in version 2.4.0 o functions bpc, bpcPca, bpcPcaTab, bpcReg, bpcRegTab, bpcTab, bpcTabWrapper added o got rid of dependencies to packages car and sjmisc Changes in version 2.3.4 o lmCoDaX now can use external non-compositional predictors as well Changes in version 2.3.3 o New data set landcover Changes in version 2.3.2 o impPCA to impute missing values with principal component analysis o outCoDa now has a method for high-dimensional data o weighted pivot coordinates Changes in version 2.3.1 (2021-09-20) o weighted pivot coordinates included o cut dependency to sRoc Changes in version 2.3.0 (2020-11-18) o new method for compositional splines added Changes in version 2.2.3 o compositional splines added Changes in version 2.2.2 o outCoDa() now contains a parameter alpha instead of h o bic is now calculated in clustCoDa Changes in version 2.2.1 (2020-02-11) o encoding issues in data sets nutrients and nutrients_branded Changes in version 2.2.0 (2020-01-14) o spline density estimation, various functions Changes in version 2.1.2 o imputeUDL modified a few parts Changes in version 2.1.1 o function variation has a new parameter method (and a new method) Changes in version 2.1.0 (2019-04-15) o new functions cubeCoord, cubeCoordWrapper, tabCoord, tabCoordWrapper o new data set employment2 o new data set manu_abs Changes in version 2.0.10 (2019-02-19) o updated documentation o going back to old version of imputeBDLs o mcd based robust variation (instead of mad) in function variation Changes in version 2.0.9 (2018-11-08) o new method in imputeBDLs o minor change in imputeBDLs o new function imputeUDLs Changes in version 2.0.8 (2018-05-29) o old Rd files management to meet new CRAN policy Changes in version 2.0.7 (2018-04-08) o bug in daFisher resolved o additional argment in missPattern o imputeBDLs now includes a new method Changes in version 2.0.6 (2017-08-14) o new method for imputation of rounded zeros in imputeBDLs Changes in version 2.0.5 (2017-07-12) o improved documentation o version of robCompositions fits now to the Springer book o orthbasis now times -1 o ilr.2x2 now fits to the formulas in the book, and not for the paper o bug in daFisher resolved o bug in external variables in pcaCoDa solved o registering C-code Changes in version 2.0.4 o aDist now again works for vectors o added following data sets: socExp, cancerMN, educFM, electionATbp, employment, employment_df, GDPsatis, govexp, instw, isic32, mortality_tab, precipitation, rcodes, teachingStuff, unemployed o corrected documentation for ageCatWorld data o new functions: coord, gmean, gmean_sum, ilr.2x2, ind2x2, indTab, int2x2, indTab, int2x2, intArray, intTab, pTab, rSDev, rSDev.test.R, SDev, stats o bug in daFisher solved and new adapted predict function Changes in version 2.0.3 (2017-01-11) o new function perturbation() and powering() o new function for inner product (iprod()) o new funciton balances() o addLR, cenLR with optional different base in logrithm o isomLR with more possibilites for a normalizing constant o isomLR depricated, using now pivotCoord instead o isomLRinv depricated, using pivotCoordInv instead o orthbasis() now returns also the basis vectors o bug in aDist solved o aDist now much faster o new function center() o method symm (symmetric balances) in pivotCoord Changes in version 2.0.2 (2016-08-22) o new function balZav() o new function daCoDa() o new function clustCoDa() o new function corCoDa() o aDist for single matrix o pcaCoDa() with external variables o new data set alcoholreg o new data set alcohol o new data set econom o new data set ageCatWorld o new data set election o new data set laborForce o new data set payments Changes in version 2.0.1 o discriminant analysis (daFisher) for n-group case o discriminant analysis (daFisher) including predict o discriminant analysis (daFisher) including misclassification rates Changes in version 2.0.0 (2016-02-08) o combined help for some S3methods o function robVariation is now called variation o new data sets Changes in version 1.9.1 (2015-05-06) o now with roxygen2 documentation o minor modifications in impRZilr Changes in version 1.9.0 (2014-08-18) o bug in robVariations solved Changes in version 1.8.0 (2014-03-16) o critical bug in function impRZilr solved (did worked only correct for constant sum data) Changes in version 1.7.0 (2014-02-07) o bug in function impRZilr solved o new bootstrap procedure for estimating the number of components within impRZilr o description in function impRZilr adapted Changes in version 1.6.4 (2013-11-26) o deprecated functions alr, ilr, clr, invalr, invilr, invclr deleted o bugfix and enhancement of impRZilr.R o improvement of isomLR for high-dimensional data Changes in version 1.6.3 (2013-02-27) o function impAll included. o robGUI excluded (is now put in the compositionsGUI package) Changes in version 1.6.2 (2013-01-11) o alr, clr, ilr, invalr, invclr, invilr are now deprecated. The new functions are called addLR, cenLR, isomLR, addLRinv, cenLRinv, isomLRinv o new functions gm, orthbasis, missPatterns and zeroPatterns o return instead of invisible return for all log-ratio transformations Changes in version 1.5.0 (2011-05-06) o variable names in expendiure data corrected o graphical user interface added o vignette about an overview of the package added o alrEM() function now renamed in impRZalr() to be consistent with notation Changes in version 1.4.6 o correct transformation of the detection limit in impRZilr when having not constant sum 1. Changes in version 1.4.5 o Bug in alrEM fixed. o impRZilr included, but without documentation. Changes in version 1.4.4 (2011-01-11) o Description for the Expenditures data changed. Wrong variable description fixed. o Wrong formula in robVariation method robust=FALSE fixed after an comment from from Mark Engle (thanks!). o Improved code in 'CITATION' file.