The fdakmapp package provides the kmap function that jointly performs clustering and alignment of a functional dataset (multidimensional or unidimensional). The centers can be computed by mean or medoid center methods. The parallel version is available.
The source code can be cloned or downloaded directly from github. An R studio project file is provided to open the project in RStudio.
The package is linked against OpenMP, the BLAS and LAPACK libraries in Makevars.
To install the package locally the packages Rcpp and RcppArmadillo needs to be installed. Using install_github() will install the dependencies (Rcpp and RcppArmadillo) automatically.
The package can be installed directly from github but devtools is required.
If devtools is not installed use the following comand to install it
install.packages('devtools')
and then install the package
library(devtools)
install_github('zitale/fdakmapp')
Usually the clag compiler doesn't support OpenMP. For this reason is available a release without -fopenmp.
library(devtools)
install_github('zitale/[email protected]')
Othewise the problem can be solved following the tutorial at: https://clang-omp.github.io/.
devtools::test()
res<-kmap(x=aneurisk65$x, y=aneurisk65$y, n_clust=2)
kmap_show_results(res,FALSE,FALSE)
The R documentation can be found in the main directory in fdakmapp.pdf. The C++ documentation can be found at https://zitale.github.io/fdakmapp/.