Multivariate Symmetric Uncertainty and Other Measurements

Estimators for multivariate symmetrical uncertainty based on the work of Gustavo Sosa et al. (2016) , total correlation, information gain and symmetrical uncertainty of categorical variables.

msu

Overview

msu is a library providing an estimator for the multivariate symmetrical uncertainty measurement and other related estimators, such as symmetrical uncertainty, total correlation, information gain (also called mutual information) and Shannon entropy for categorical variables.

For more details about MSU read Understanding a Version of Multivariate Symmetric Uncertainty to assist in Feature Selection and references therein.

Installation

To install msu from CRAN

install.packages("msu")

The development version is available in GitHub and it can be installed using devtools

devtools::install_github("eliasmacielr/msu")