Cluster Analysis: An Application to a Real Mixed-Type Data Set

Abstract

When you dispose of multivariate data it is crucial to summarize them, so as to extract appropriate and useful information, and consequently, to make proper decisions accordingly. Cluster analysis fully meets this requirement; it groups data into meaningful groups such that both the similarity within a cluster and the dissimilarity between groups are maximized. Thanks to its great usefulness, clustering is used in a broad variety of contexts; this explains its huge appeal in many disciplines. Most of the existing clustering approaches are limited to numerical or categorical data only. However, since data sets composed of mixed types of attributes are very common in real life applications, it is absolutely worth to perform clustering on them. In this paper therefore we stress the importance of this approach, by implementing an application on a real world mixed-type data set.