Geometric Data Analysis and related techniques
GDAtools provides functions for Geometric Data Analysis :
- Specific Multiple Correspondence Analysis
- Class Specific Analysis
- Nonsymmetric Correspondence Analysis
- two-table and k-table analyses : Multiple Factor Analysis, between-
and within-class analysis, MCA and PCA with instrumental variables or
orthogonal instrumental variables, discriminant analysis, coinertia
analysis, etc.
- guides for interpretation (contributions, quality of representation,
test-values, etc.)
- analysis of structuring factors (concentration ellipses,
interactions, etc.)
- inductive analysis (typicality and homogeneity tests, confidence
ellipses)
- bootstrap validation
- many graphical representations for MCA and variants (with and
without ggplot2)
- plots for hierarchical clustering
Available on github and CRAN.
Please visit https://nicolas-robette.github.io/GDAtools/ for
documentation.
R package descriptio
Descriptive Statistical Analysis
descriptio provides functions for the description of statistical
associations between two variables :
- measures of local and global association between variables (phi,
Cramer’s V, point-biserial correlation, eta-squared, Goodman &
Kruskal tau, PEM, etc.),
- graphical representations of the associations between two variables
(using ggplot2),
- weighted statistics,
- permutation tests.
Available on github and CRAN.
Please visit https://nicolas-robette.github.io/descriptio/ for
documentation.
R package moreparty
Tools for conditional inference random forests
This package aims at complementing the party package with
parallelization and interpretation tools.
It provides functions for :
- parallelized conditional random forest
- parallelized variable importance
- feature selection : recursive and non-recursive feature elimination,
algorithms based on permutation tests
- accumulated local effects (ALE), partial dependence and interaction
strength
- surrogate tree
- prototypes
- getting any tree from a forest
- assessing the stability of a conditional tree
- bivariate association measures
- dot plots for variable importance and effects
Available on github and CRAN.
Please visit https://nicolas-robette.github.io/moreparty/ for
documentation.
R package seqhandbook
a companion package for my handbook on sequence
analysis
It provides the datasets used in the examples in the handbook, as
well as functions for :
- describing episodes in individual sequences (at least one episode,
number of episodes, position of the start of the first episode)
- measuring association between domains in multidimensional sequence
analysis
- heat maps of sequence data
- Globally Interdependent Multidimensional Sequence Analysis
(GIMSA)
- smoothing sequences for index plots
- coding sequences for Qualitative Harmonic Analysis
- measuring stress from MDS factors
- symmetrical PLS
Available on github and CRAN.
Please visit https://cran.r-project.org/web/packages/seqhandbook/vignettes/Tutoriel.html
for documentation.