Абстрактный

High Quality Assessment of Similarity by Using Multiple View Points

K.Ramesh, C.Vasumurthy,Prof.D.Venkatesh

Clustering is a process of grouping objects based on certain similarity measure. Then groups are known as clusters which can be analyzed and used further for operations like query processing. Clustering algorithms assume certain relationship among objects in the given dataset. The existing clustering algorithms with respect to text mining use single viewpoint similarity measure for partitioned clustering of objects. The main drawback of these algorithms is that the resultant clusters can’t make use of fully informative assessment. In this paper we propose a new measure for finding similarity between objects which is multi-viewpoint based. This approach considers multiple viewpoints while comparing objects for clustering. This measure can have more informative assessment of similarity thus making clusters with highest quality. We also proposed two criterion approaches for achieving highest intra-cluster similarity and lowest inter-cluster similarity. The empirical results reveal that the proposed measure is used in making quality clusters.