Optimizing Clustering Technique based on Partitioning DBSCAN and Ant Clustering Algorithm
Chaudhari Chaitali G.
Chaitali Chaitali, Department of Information Technology Parul Institute of Engg. Technological University, Gujarat, India.
Manuscript received on November 24, 2012. | Revised Manuscript received on December 11, 2012. | Manuscript published on December 30, 2012. | PP: 212-215 | Volume-2, Issue-2, December 2012. | Retrieval Number: B0900112212 /2012©BEIESP
Open Access | Ethics and Policies | Cite
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Clustering is the process of organizing similar objects into the same clusters and dissimilar objects in to different cluster. Similarities between objects are evaluated by using the attribute value of object, a distance metric is used for evaluating dissimilarity. DBSCAN algorithm is attractive because it can find arbitrary shaped clusters with noisy outlier and require only two input parameters. DBSCAN algorithm is very effective for analyzing large and complex spatial databases. DBSCAN need large volume of memory support and has difficulty with high dimensional data. Partitioning-based DBSCAN was proposed to overcome these problems. But both DBSCAN and PDBSCAN algorithms are sensitive to the initial parameters.
Keywords: Clustering, DBSCAN, PDBSCAN, Ant clustering algorithm.