Density Based Clustering Scheme using Dynamic Dissimilarity Measures
V. Kavitha1, R. Manavalan2
1Kavitha V, Computer Science, Periyar University, KSR College of Arts and Science ,Tiruchengode.
2Manavalan, Computer Science, Periyar University, KSR College of Arts and Science, Tiruchengode.
Manuscript received on March 19, 2012. | Revised Manuscript received on April 22, 2012. | Manuscript published on April 30, 2012. | PP: 101-107 | Volume-1 Issue-4, April 2012 | Retrieval Number: D0289041412/2012©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 methods are used support estimates of a data distribution have newly attracted much attention because of their ability to generate cluster boundaries of arbitrary shape and to contract with outliers efficiently. This paper proposes, Density based clustering using dynamic dissimilarity measure based on a dynamical system associated with Density estimating functions. Hypothetical basics of the proposed measure are developed and applied to construct a clustering method that can efficiently partition the whole data space. Clustering based on the proposed dissimilarity measure is robust to handle large amount of data and able to estimate the number of clusters automatically by avoid overlap. The dissimilarity values are evaluated and clustering process is carried out with the density values.
Keywords: Clustering, kernel methods, dynamical systems, equilibrium vector, support, density.