Variables Clustering

Performs clustering of quantitative variables, assuming that clusters lie in low-dimensional subspaces. Segmentation of variables, number of clusters and their dimensions are selected based on BIC. Candidate models are identified based on many runs of K-means algorithm with different random initializations of cluster centers.

README

What is this repository for?

Package varclust

• is R package for clustering quantitative variables
• provides estimation of number of clusters
• enables significant data dimension reduction

Check out demo version!

How do I get set up?

• Install varclust package using devtools package

install_github("psobczyk/varclust")

• Download the package as an archive and install it manually from R console

• You might need to install package dependencies:

  • parallel
  • doParallel
  • foreach
  • doRNG
  • RcppEigen
  • pesel

• The pesel package can be installed using devtools package

install_github("psobczyk/pesel")

• No additional configuration is needed
• Read vignette to get familiar with basic usage

Who do I talk to?

• If help provided in the package documentation does not solve your problem please contact Piotr.Sobczyk[at]pwr.edu.pl

This project has received funding from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement no 602552.