Skip to content

coda.base 1.0.7 (Release date: 2026-05-12)

  • dist_coda() now accepts method = "L2" as an alias for the Euclidean distance.
  • The partial principal-balance search validates grouped inputs more thoroughly and improves the handling of tabu-search neighbourhoods.
  • Additional tests cover distance aliases, basis construction and partial tabu-search edge cases.
  • The pkgdown site now uses Bootstrap 5 and includes a redesigned logo, favicons and a revised Get Started vignette.

coda.base 1.0.6 (Release date: 2026-05-08)

CRAN release: 2026-05-08

coda.base 1.0.5 (Release date: 2026-03-03)

CRAN release: 2026-03-04

  • Reworked the basis and coordinate code, with stronger checks for dimensions, names and invalid compositions.
  • Added cdp_basis() and conditional_obasis() to the public API.
  • Added gen_coda_with_zeros_and_missings() for generating data sets used to assess zero and missing-value replacement methods.
  • Improved coda_replacement() and the conditional orthonormal-basis calculations used for zero and missing-value imputation.
  • Improved sbp_basis(), constrained principal balances and the C++ principal-balance implementation.
  • plot_balance() can display statistics associated with balances.
  • Optimised and reorganised the C++ code and expanded the basis tests.

coda.base 1.0.4 (Release date: 2025-09-20)

  • Changed the sign convention of CoDaPack bases to agree with CoDaPack.

coda.base 1.0.3 (Release date: 2025-07-02)

CRAN release: 2025-07-02

  • Improved completion of sequential binary partitions in sbp_basis().
  • Improved constrained principal-balance calculations and their handling of subcompositions.

coda.base 1.0.2 (Release date: 2025-05-07)

  • Added the kilauea_iki compositional data set and its documentation.
  • Improved documentation for zero replacement and coordinate transformations.
  • Improved dimension checks for pairwise bases and composition reconstruction.

coda.base 1.0.0 (Release date: 2025-05-02)

CRAN release: 2025-05-02

  • Added coda_replacement() for imputing rounded zeros and missing values in compositional data.
  • Added conditional orthonormal-basis calculations used by the imputation procedure.
  • Reorganised basis, coordinate and constrained principal-balance code, with more consistent argument validation and naming.
  • Updated the package documentation and test suite for the 1.0 release.

coda.base 0.5.5 (Release date: 2023-11-25)

CRAN release: 2023-11-25

coda.base 0.5.4 (Release date: 2023-03-01)

  • Added a collection of documented compositional data sets, including alimentation, blood_mn, bmi_activity, eurostat_employment, foraminiferals, house_expend, mammals_milk, milk_cows, montana, petrafm, pollen, serprot, statistitian_time and weibo_hotels.
  • plot_balance() now uses balance names in dendrogram labels.
  • Fixed completion of incomplete sequential binary partitions in sbp_basis().
  • Removed cc_basis() because its results did not match the intended method; it was redesigned and restored in a later release.

coda.base 0.5.3 (Release date: 2023-02-08)

  • sbp_basis() can complete an incomplete sequential binary partition to obtain a full basis.

coda.base 0.5.2 (Release date: 2022-07-14)

CRAN release: 2022-07-18

  • Basis constructors consistently return standard R matrices; sparse matrices are used internally where they improve performance.
  • Updated the minimum RcppArmadillo requirement for the compiled code.

coda.base 0.5.1 (Release date: 2022-07-07)

  • Added the Wilks statistic as a criterion for canonical-correlation balances.

coda.base 0.4.3

  • center() and gmean() functions incorporated.

coda.base 0.4.2

  • Pairwise default name changed to lr instead of alr.
  • Sparse matrices are used for logcontrast matrices of balances.
  • Problem with constrained principal balances detected and solved.

coda.base 0.3.2 (Release date: 2022-03-13)

  • coord() and comp() are included as a shorter names of coordinates() and composition().
  • Basis names taken from composition.
  • Coordinates based on pairwise log-ratios available with pairwise_basis() or basis = ‘pw’ in function coordinates().
  • Parameter label removed in function coordinates().

coda.base 0.3.1

CRAN release: 2020-05-14

  • PB constrained approximation available. Adaptation of code provided by J.A.Martín-Fernández.
  • PB exact algorithm rewritten to improve performance.
  • PB local search algorithm removed.

coda.base 0.2.2

CRAN release: 2020-04-16

  • MASS dependence removed
  • Code reorganised to avoid repetition.

coda.base 0.2.1

CRAN release: 2019-07-02

  • ?_basis functions always return a matrix
  • consistencies between methods improved

coda.base 0.1.12 (Release date: 2018-04-15)

CRAN release: 2019-04-15

  • coordinates function keeps row.names attribute
  • Principal components calculations are done using svd function which is faster than princomp
  • sbp_basis function does not need package to be loaded, can run as coda.base::sbp_basis()

coda.base 0.1.11 (Release date: 2018-12-23)

CRAN release: 2019-01-15

  • Principal components coordinates uncentered.
  • Sequential binary partition can be defined using a matrix with -1,0 and 1’s.
  • CoDaPack’s basis included.
  • A vignette explaining how to build coordinates has been included.

coda.base 0.1.10 (Release date: 2018-08-03)

CRAN release: 2018-08-06

  • Principal Balances included as direct method in coordinates function.
  • Coordinates function filters non-valid compositions.

coda.base 0.1.9 (Release date: 2018-05-03)

CRAN release: 2018-05-09

  • Modifications included to rcran policies.

coda.base 0.1.8 (Release date: 2017-12-03)

CRAN release: 2017-12-03

  • Exact algorithm for Principal Balances included.
  • Bug in lsearch base in principal components solved.

coda.base 0.1.7 (Release date: 2017-08-25)

CRAN release: 2017-09-26

  • Problem with function composition using ‘clr’ basis solved.
  • ‘coda.base.basis’ option included to hide the basis when printing the coordinates.
  • New method which uses the principal components included in pb_basis.
  • Documentaion improved.