| Abstract : | The need for obtaining groups of objects that have natural and useful properties arises in many practical applications. The very fact that usually no grouping being known a priori exists, implies that cluster analysis is a fundamental and powerful tool for organizing and investigating possible relations within multivariate data. Using the information embedded in some variables, cluster analysis intends to partition the whole set of observations into individual sets of similar characteristics. Although several algorithmic processes have been proposed for clustering, the model based method gains ground ever more. On the one hand, it is the opportunity it offers for evaluating the findings and for statistical inference; on the other hand, it is the increasing interest for investigating distributions that better fit real data, as well as the development of computer systems that have set the above method popular not only within Statistics, but also between fields of other sciences, such as the Archaeometry.
|
|---|
Copyright © 2013 Library AUEB. All rights reserved.