Proposed KDBSCAN Algorithm for Clustering

  • Yossra Hussein Department of Computer Science, College of Science, Technology University, Baghdad, Iraq
  • Safa Abdel Jalil Department of Computer Science, College of Science, Technology University, Baghdad, Iraq

Abstract

Science, technology and many other fields are use clustering algorithm widely for many applications, this paper presents a new hybrid algorithm called KDBSCAN that work on improving k-mean algorithm and solve two of its
problems, the first problem is number of cluster, when it`s must be entered by user, this problem solved by using DBSCAN algorithm for estimating number of cluster, and the second problem is randomly initial centroid problem that has been dealt with by choosing the centroid in steady method and removing randomly choosing for a better results, this work used DUC 2002 dataset to obtain the results of KDBSCAN algorithm, it`s work in many application fields such as electronics libraries, biology and marketing, the KDBSCAN algorithm that described in this paper has better results than traditional K-mean and DBSCAN algorithms in many aspects, its
preform stable result with lower entropy.

Published
Jan 30, 2018
How to Cite
HUSSEIN, Yossra; ABDEL JALIL, Safa. Proposed KDBSCAN Algorithm for Clustering. Iraqi Journal of Science, [S.l.], p. 173-178, jan. 2018. ISSN 2312-1637. Available at: <http://scbaghdad.edu.iq/eijs/index.php/eijs/article/view/47>. Date accessed: 19 aug. 2018.
Section
Computer Science