文章基本信息

标题：Hybrid Algorithm for Noise-free High Density Clusters with Self-Detection of Best Number of Clusters
本地全文：下载
作者：T. Karthikeyan ; S. John Peter ; S.Chidambaranathan 等
期刊名称：International Journal of Hybrid Information Technology
印刷版ISSN：1738-9968
出版年度：2011
卷号：4
期号：2
出版社：SERSC
摘要：Clustering is a process of discovering group of objects such that the objects of the samegroup are similar, and objects belonging to different groups are dissimilar. A number ofclustering algorithms exist that can solve the problem of clustering, but most of them are verysensitive to their input parameters. Minimum Spanning Tree clustering algorithm is capableof detecting clusters with irregular boundaries. A density based notion of clusters which isdesigned to discover clusters of arbitrary shape. In this paper we propose a combinedapproach based on Minimum Spanning Tree based clustering and Density based clusteringfor noise free high density best number of clusters. The algorithm uses a new clustervalidation criterion based on the geometric property of data partition of the data set in orderto find the proper number of clusters at each level. The algorithm works in two phases. Thefirst phase of the algorithm produces subtrees (noise free clusters). The second phase findshigh density clusters from the subtrees
关键词：Center; Clustering; Cluster validity; Cluster Separation; Euclidean minimum;spanning tree; Eccentricity; Outliers; Subtree