期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2011
卷号：2
期号：2
页码：607-610
出版社：TechScience Publications
摘要：One of the major problems in cluster analysis is the determination of the number of clusters in unlabeled data, which is a basic input for most clustering algorithms. Typically, the clustering algorithm partitions a dataset into a fixed number of clusters supplied by the user, i.e., Given a dataset O representing n Objects {o1, o2… on}, clustering aims to partitions data into c groups, i.e., C1, . . . . Cc, so that Ci ∩ Cj = Ø and C1 Ù C2 Ù C3 Ù …….Ù Cc =O. The present paper propose a novel method ,which is based on Layered Hidden Markov Model(LHMM) to identify a suitable number of clusters in a given unlabeled dataset without using prior knowledge about the number of clusters. For this, the present paper partitions the dataset into windows of fixed/different size based on a novel scheme called log likelihood values of HMM. The proposed scheme works as a framework for identifying the appropriate number of clusters. The proposed method is implemented on Iris dataset. The experimental results indicate the efficacy of the proposed method.
