Document: multimix-ppaper
Title: Mixture Model Clustering with the Multimix Program
Author: Murray Jorgensen and Lynette Hunt
Abstract: Hunt has implemented the finite mixture model approach to 
 clustering in a program called Multimix.  The program is designed to
 cluster multivariate data with categorical and continuous variables
 and possibly containing missing values.  In this presentation we
 describe the approach taken to the design of Multimix and how some of
 the statistical problems were dealt with.  As examples of the use of
 the program we cluster a large medical dataset and a version of
 Fisher's Iris data in which a third of the values are randomly made
 `missing'.
Section: Science/Data Analysis

Format: PDF
Files: /usr/share/doc/multimix/PPAPER.pdf