Model based cluster analysis in data mining

Classical analysis methods do not yield efficient results when applied to large scale datasets. For this reason, data mining methods have been developed through collective studies of various disciplines. One of the main purposes of data mining is to identify the unknown group structure on the dataset. Cluster analysis is the general name of methods that aim to determine the grouping on the dataset. One of these methods, which is model-based cluster analysis, is clustering dataset using finite mixture models. With the use of finite mixture models in the cluster analysis, the problem of cluster determination is transformed into statistical modeling problem. In this thesis, model-based clustering analysis is examined. Data mining applications of this method have been implemented and compared with other methods.