kamila (

Methods for Clustering Mixed-Type Data.


Implements methods for clustering mixed-type data, specifically combinations of continuous and nominal data. Special attention is paid to the often-overlooked problem of equitably balancing the contribution of the continuous and categorical variables. This package implements KAMILA clustering, a novel method for clustering mixed-type data in the spirit of k-means clustering. It does not require dummy coding of variables, and is efficient enough to scale to rather large data sets. Also implemented is Modha-Spangler clustering, which uses a brute-force strategy to maximize the cluster separation simultaneously in the continuous and categorical variables.

Maintainer: Alexander Foss
Author(s): Alexander Foss [aut, cre], Marianthi Markatou [aut]

License: GPL-3 | file LICENSE

Uses: abind, gtools, KernSmooth, mclust, plyr, Rcpp, Hmisc, ggplot2, testthat, clustMD

Released about 2 years ago.