New function

  • scaled.dev() : scaled deviations between categories of a supplementary categorical variable

Bug fixes

  • ggadd_supind() : bug fix when only one supplementary individual (thanks to Mathieu Ferry)

New function

Changes in existing functions

Tiny fixes

  • dimdescr() : fixed column names in the results + junk categories are not displayed for speMCA results
  • ggadd_density() : fixed deprecated ggplot2 arguments
  • bcMCA() : bug fix when there are junk categories

Changes in existing functions

Bug fixes

Please note that the 1.8 version of GDAtools was not published on CRAN. So, compared to the last version on CRAN, 2.0 version inherits the changes of 1.8 version.

Major change

New functions

  • gPCA() : Generalized Principal Component Analysis
  • bcMCA() : Between-class Multiple Correspondence Analysis
  • bcPCA() : Between-class Principal Component Analysis
  • wcMCA() : Within-class Multiple Correspondence Analysis
  • wcPCA() : Within-class Principal Component Analysis
  • PCAiv() : Principal Component Analysis with Instrumental Variables
  • MCAiv() : Multiple Correspondence Analysis with Instrumental Variables
  • PCAoiv() : Principal Component Analysis with Orthogonal Instrumental Variables
  • MCAoiv() : Multiple Correspondence Analysis with Orthogonal Instrumental Variables
  • coiPCA() : Coinertia analysis between two groups of numerical variables
  • coiMCA() : Coinertia analysis between two groups of categorical variables
  • DA() : Descriptive Discriminant Analysis
  • DAQ() : Descriptive Discriminant Analysis with Qualitative Variables (aka disqual)
  • rvcoef() : RV coefficient between two groups of variables
  • planecontrib() : For a given plane of a MCA, computes contributions et squared cosines of the active variables and categoriesand of the individuals
  • ggeta2_variables() : Plots the eta-squared of the active variables of a MCA
  • quasindep() : Transforms a symmetrical contingency table so that it can be used for quasi-correspondence analysis, also called correspondence analysis of incomplete contingency table
  • ggsmoothed_supvar() : Plots the density a supplementary variable in a MCA space
  • bootvalid_variables() : Bootstrap validation for active variables of a MCA
  • bootvalid_supvars() : Bootstrap validation for supplementary variables of a MCA
  • ggbootvalid_variables() : Ellipses for bootstrap validation of active variables of a MCA
  • ggbootvalid_supvars() : Ellipses for bootstrap validation of supplementary variables of a MCA
  • supind() : replaces indsup(), which is softly deprecated
  • supvar() : replaces varsup(), which is softly deprecated
  • supvars() : replaces varsups(), which is softly deprecated
  • nsCA() : Nonsymmetric Correspondence Analysis

Changes in existing functions

Other changes

  • The package has much fewer dependencies : some packages are no longer needed, other have been moved to Suggests.

New functions

  • ijunk(): Shiny app to select interactively the junk categories before a specific MCA.
  • quadrant(): Computes the quadrant of active individuals in a given space of a MCA.
  • oddsratio.table(): Computes the odds ratio for every cell in a contingency table.
  • ggassoc_chiasmogram(): Plots the chiasmogram of a crosstabulation, using ggplot2.
  • ggassoc_assocplot(): Association plot of a crosstabulation, using ggplot2.
  • ggassoc_bertin(): Bertin plot of a crosstabulation, using ggplot2.
  • ahc.plots(): Various plots of Ascending Hierarchical Clustering.
  • dist.chi2(): Computes chi-squared distance.
  • ggaxis_variables(): Plots variables on a single axis of a MCA.
  • varsups(): Computes statistics for categorical supplementary variables.
  • ggadd_supvars(): Adds categorical supplementary variables to a cloud of variables.

Bug fixes

  • bug fix in speMCA(), csMCA() and getindexcat() when empty levels or non-factor vectors in the data
  • bug fix in indsup() : supdata can now be a tibble
  • bug fix in assoc.yx() : integers are now allowed for y; empty levels are dropped in x
  • bug fix in wtable() : empty cells are replaced by 0.

Changes in existing functions

  • speMCA() and csMCA() : junk categories can now be specified as a character vector
  • csMCA() : results can now be used with explor package
  • tabcontrib() : new “best” option (thanks to @419kfj)
  • assoc.twocat() : standardized (i.e. Pearson) residuals, adjusted standardized residuals, odds ratios, PEM and Goodman-Kruskal tau are computed. The object is reorganized into several sublists. “gather” data frame has columns for margins frequencies and percentages.
  • ggassoc_crosstab() : rewriting with several new options (size, measure, limit, palette and direction) and no more dependency to GGally package
  • ggassoc_phiplot(), ggassoc_assocplot() and ggassoc_crosstab() : now allow faceting. The measure of local association can be any one computed by assoc.twocat()
  • ggadd_interaction() : geom_line replaced by geom_path (thanks to @419kfj)
  • ggadd_chulls() : new “prop” option to allow peeling of the hull

Other

  • Removed dependency to GGally package

Bug fixes

New functions

  • angles.csa(): Computes the cosines similarities and angles between the dimensions of a CSA and those of a MCA.

Bug fixes

Changes in existing functions

New functions

  • ggadd_density(): adds a density layer to the cloud of individuals for a category of a supplementary variable
  • ggadd_corr(): adds a heatmap of under/over-representation of a supplementary variable to a cloud of individuals
  • ggadd_kellipses() : adds concentration ellipses to a cloud of individuals, using ggplot
  • ggadd_chulls() : adds convex hulls to a cloud of individuals, using ggplot
  • ggassoc_crosstab() : plots counts and associations of a crosstabulation, using ggplot
  • ggassoc_phiplot() : bar plot of phi measures of association of a crosstabulation, using ggplot
  • ggassoc_boxplot() : displays of boxplot and combines it with a violin plot, using ggplot
  • ggassoc_scatter() : scatter plot with a smoothing line, using ggplot
  • dimdescr() : works with condesc() instead of FactoMineR::condes() and takes row weights into account.
  • dimtypicality() : computes typicality tests for supplementary variables
  • ggadd_attractions() : adds attractions between categories (via segments) to a cloud of variables
  • ggadd_supind() : adds supplementary individuals to a cloud of individuals, using ggplot
  • flip.mca() : flips the coordinates of the individuals and the categories on one or more dimensions of a MCA

Removed functions :

Changes in existing functions

  • ggcloud_indiv() : the density of points can be represented as an additional layer through contours or hexagon bins
  • catdesc() and condesc() : allow weights
  • catdesc() and condesc() : new nperm and distrib options
  • catdesc() and condesc() : new robust option
  • assoc.twocont(), assoc.twocat() and assoc.catcont() : nperm option is set to NULL by default
  • darma() : nperm is set to 100 by default
  • ggcloud_variables() and ggcloud_indiv() : a few changes in the theme (grids are removed, etc.)
  • ggcloud_indiv() and ggadd_ellipses() : new size option
  • ggcloud_variables() : new min.ctr option to filter categories according to their contribution (for objects of class MCA, speMCA and csMCA)
  • ggcloud_variables() : new max.pval option to filter categories according to the p-value derived from their test-value (for objects of class stMCA and multiMCA)
  • ggcloud_variables() : prop argument can take values “vtest1” and “vtest2”
  • ggcloud_variables() : for shapes and colors, variables are used in their order of appearance in the data instead of alphabetical order
  • ggcloud_variables() : new face argument to use font face to identify the most contributing categories
  • homog.test() : gives the p-values in addition to the test statistics
  • dimeta2() : l argument renamed to vars and n argument removed
  • varsup() : also computes typicality tests and correlation coefficients
  • conc.ellipse() : several kinds of inertia ellipses can be plotted thanks to the kappa option
  • ggadd_ellipses() : level is set to 0.05 by default, which corresponds to conventional confidence ellipses. Option ‘points’ to choose to color the points or not.
  • modif.rate() : computes raw and modified rates
  • homog.test() : new dim argument
  • modif.rate() : compatibility with objects of class MCA, speMCA, csMCA, stMCA and multiMCA
  • ggcloud_variables() : compatibility with objects of class MCA, speMCA, csMCA, stMCA and multiMCA
  • ggcloud_indiv() : compatibility with objects of class MCA, speMCA, csMCA, stMCA and multiMCA
  • ggadd_supvar() : compatibility with objects of class MCA, speMCA, csMCA, stMCA and multiMCA
  • ggadd_interaction() : compatibility with objects of class MCA, speMCA, csMCA, stMCA and multiMCA
  • dimeta2() : compatibility with objects of class MCA, speMCA, csMCA, stMCA and multiMCA
  • dimcontrib() : compatibility with objects of class MCA, speMCA and csMCA
  • tabcontrib() : compatibility with objects of class MCA, speMCA and csMCA
  • homog.test() : compatibility with objects of class MCA, speMCA, csMCA, stMCA and multiMCA
  • varsup() : compatibility with objects of class MCA, speMCA, csMCA, stMCA and multiMCA
  • ggadd_chulls() : compatibility with objects of class MCA, speMCA, csMCA, stMCA and multiMCA
  • ggadd_corr() : compatibility with objects of class MCA, speMCA, csMCA, stMCA and multiMCA
  • ggadd_density() : compatibility with objects of class MCA, speMCA, csMCA, stMCA and multiMCA
  • ggadd_ellipses() : compatibility with objects of class MCA, speMCA, csMCA, stMCA and multiMCA
  • ggadd_kellipses() : compatibility with objects of class MCA, speMCA, csMCA, stMCA and multiMCA

Bug fixes

New functions

  • phi.table() : computes phi coefficient for every cells of a contingency table
  • assoc.twocont() : measures the association between two continuous variables with Pearson, Spearman and Kendall correlations and a permutation test.
  • assoc.yx() : computes bivariate association measures between a response and predictor variables
  • darma() : computes bivariate association measures between a response and predictor variables, displaying results in a form looking like the summary of a regression model analysis.

Bug fixes

Changes in existing functions

New functions

  • assoc.twocat(): measures the association between two categorical variables
  • assoc.catcont(): measures the association between a categorical variable and a continuous variable
  • catdesc(): measures the association between a categorical variable and some continuous and/or categorical variables
  • condesc(): measures the association between a continuous variable and some continuous and/or categorical variables
  • ggcloud_indiv(): cloud of individuals, using ggplot
  • ggcloud_variables(): cloud of variables, using ggplot
  • ggadd_supvar(): adds a supplementary variable to a cloud of variables, using ggplot
  • ggadd_interaction(): adds the interaction between two variables to a cloud of variables, using ggplot
  • ggadd_ellipses(): adds confidence ellipses to a cloud of individuals, using ggplot

Changes in existing functions

  • conc.ellipses(): additional options

New functions

  • translate.logit(): translates logit models coefficients into percentages
  • tabcontrib(): displays the categories contributing most to MCA dimensions

Changes in existing functions

  • varsup(): with csMCA, the length of variable argument can be equal to the size of the cloud or the subcloud
  • textvarsup(): with csMCA, the length of variable argument can be equal to the size of the cloud or the subcloud
  • conc.ellipse(): with csMCA, the length of variable argument can be equal to the size of the cloud or the subcloud
  • plot.multiMCA(): threshold argument, aimed at selecting the categories most associated to axes
  • plot.stMCA(): threshold argument, aimed at selecting the categories most associated to axes

Changes in existing functions

  • dimdesc.MCA(): now uses weights

Bug fixes

  • dimdesc.MCA(): problem of compatibility next to a FactoMineR update

New functions

  • dimvtest(): computes test-values for supplementary variables

Changes in existing functions

New functions

  • wtable(): works as table() but allows weights and shows NAs as default
  • prop.wtable(): works as prop.table() but allows weights and shows NAs as default

Changes in existing functions

  • multiMCA(): RV computation is now an option, with FALSE as default, which makes the function execute faster

Bug fixes

  • textvarsup(): there was an error with the supplementary variable labels when resmca was of class csMCA.

Error fixes

  • textvarsup(): plots supplementary variables on the cloud of categories (and not the cloud of individuals as it was mentioned in help).