Kernel Learning Integrative Clustering

Kernel Learning Integrative Clustering (KLIC) is an algorithm that allows to combine multiple kernels, each representing a different measure of the similarity between a set of observations. The contribution of each kernel on the final clustering is weighted according to the amount of information carried by it. As well as providing the functions required to perform the kernel-based clustering, this package also allows the user to simply give the data as input: the kernels are then built using consensus clustering. Different strategies to choose the best number of clusters are also available. For further details please see Cabassi and Kirk (2020) .

KLIC - Kernel Learning Integrative Clustering

Installing

In order to install this package, you will need to have Rmosek installed.

Installing Rmosek [Unix]

• Download mosek from https://www.mosek.com/downloads/ to any folder (usually /home/<username>/bin). We will refer to it as <my-mosek-directory>.
• Unzip it:

cd <my-mosek-directory>
tar -xvf mosektoolslinux64x86.tar.bz2

• Add the following to your ~/.bashrc file:

export PATH=$PATH:<my-mosek-directory>/mosek/8/tools/platform/linux64x86/bin

• If you are eligible apply for a free personal academic license at https://www.mosek.com/products/academic-licenses/ You will receive it (immediately) via email.
• Put the license file in <my-mosek-directory>/mosek
• Proceed with the installation of mosek

export PKG_MOSEKHOME=<my-mosek-directory>/mosek/8/tools/platform/linux64x86
export PKG_MOSEKLIB=mosek64

• Now that you have mosek installed, you can open R and install the Rmosek package with

install.packages("Rmosek", type="source", INSTALL_opts="--no-multiarch",
                  repos="http://download.mosek.com/R/8")

Installing klic

You can install the latest released version of klic from CRAN with

install.packages("klic")

or the development version from GitHub with

library(devtools)
install_github("acabassi/klic")
library(klic)