Abstract: All clustering ways ought to assume some cluster relationship among the data objects that they are applied on. Similarity between a pair of objects are often outlined either expressly or implicitly. During this paper, I have a tendency to introduce a unique multi-viewpoint primarily based similarity live and to connected bunch ways. The most important distinction between a conventional dissimilarity/similarity live and mine is that the previous uses solely one viewpoint, that is that the origin, whereas the latter utilizes many various viewpoints, that square measure objects assumed to not be within the same cluster with the two objects being measured. Exploitation multiple viewpoints, additional informative assessment of similarity may be achieved. Theoretical analysis and empirical study square measure conducted to support this claim. Two criterion functions for document bunch are projected supported this new measure. I compare them with many well-known bunch algorithms that use alternative common similarity measures on varied document collections to verify the benefits of this proposal.
Keyword: The Clustering, K –means algorithms.
Title: DOCUMENT CLUSTERING
Author: M. Supriya
International Journal of Interdisciplinary Research and Innovations
ISSN 2348-1218 (print), ISSN 2348-1226 (online)
Research Publish Journals
