Title: Model-based Clustering

Abstract: Mixture models extend the toolbox of clustering methods available to the data analyst. They allow for an explicit definition of the cluster shapes and structure within a probabilistic framework and exploit estimation and inference techniques available for statistical models in general. In this chapter an introduction to cluster analysis is provided, model-based clustering is related to standard heuristic clustering methods and an overview on different ways to specify the cluster model is given. Post-processing methods to determine a suitable clustering, infer cluster distribution characteristics and validate the cluster solution are discussed. The versatility of the model-based clustering approach is illustrated by giving an overview on the different areas of applications.