Rdimtools is an R implementation of a number of dimension reduction and estimation methods implemented using RcppArmadillo for efficient computations. Please see the section below for the complete composition of this package and what we can provide in a unifying interface across many methods.
The package consists of three families of functions whose names start with do.,est. and aux.
for performing dimension reduction/manifold learning, estimating intrinsic dimension, and some efficient
implementations of other useful methods respectively.
do. functions are for dimension reduction (or, manifold learning) methods.
A simple taxonomy of the methods would be to categorize based on the linearity of
embedding mappings. For linear methods, we have
do.fa for (Exploratory) Factor Analysis.
do.ica for Independent Component Analysis.
do.lda for Linear Discriminant Analysis.
do.lmds for Landmark Multidimensional Scaling.
do.lpp for Locality Preserving Embedding (LPP)
do.mds for Multidimensional Scaling.
do.npe for Neighborhood Preserving Embedding.
do.pca for Principal Component Analysis.
do.rndproj for Random Projection.
Also, we have nonlinear methods implemented
do.cisomap for Conformal Isometric Feature Mapping.
do.dm for Diffusion Maps.
do.isomap for Isometric Feature Mapping.
do.keca for Kernel Entropy Component Analysis.
do.kpca for Kernel Principal Component Analysis.
do.lapeig for Laplacian Eigenmaps.
do.lisomap for Landmark Isometric Feature Mapping.
do.lle for Locally Linear Embedding.
do.ltsa for Local Tangent Space Alignment.
do.mvu for Maximum Variance Unfolding / Semidefinite Embedding.
do.plp for Piecewise Laplacian Projection.
do.ree for Robust Euclidean Embedding.
do.sammon for Sammon Mapping.
do.sne for Stochastic Neighbor Embedding.
do.tsne for t-distributed Stochastic Neighbor Embedding.
Secondly, est. family of functions are for intrinsic dimension estimation methods, including
est.boxcount for Box-Counting Dimension.
est.correlation for Correlation Dimension.
Finally, there are some auxiliary functions (aux. family),
aux.gensamples to generate samples from predefined shapes.
aux.graphnbd to make a neighborhood graph given certain criteria.
aux.kernelcov that computes a centered gram matrix with 20 kernels supported.
aux.preprocess to perform preprocessing of centering, decorrelating, or whitening.
aux.shortestpath is an efficient implementation of Floyd-Warshall algorithm.