Cluster analysis

the long version

This book is designed to be an introduction to cluster analysis for those with no background and for those who need an up-to-date and systematic guide through the maze of concepts, techniques, and algorithms associated with the clustering data. The authors begin by discussing measures of similarity, the input needed to perform any clustering analysis. They note varying theoretical meanings of the concept and discuss the set of empirical measures most commonly used to measure similarity. Various methods for actually identifying the clusters are then described. Finally, they discuss procedures for validating the adequacy of a cluster analysis. At all points, the differing concepts and techniques are compared and evaluated.