Data mining: Clustering

Abstract

This article presents a broad overview of the main clustering methodologies. It is accomplished by introducing the clustering problem and the key elements characterizing it. In particular, we describe different distance and similarity measures which can be used in a clustering method. Then, we introduce a categorization of the clustering methods and describe some relevant algorithms belonging to each category. In order to contextualize the presented methods, we provide a section reporting some relevant application cases of the clustering algorithms in different domains. Finally, we discuss about measures and criteria which have been commonly adopted for clustering evaluation.