Improved search strategies and extensions to k-medoids-based clustering algorithms

  • Authors:
  • Shu-Chuan Chu;John F. Roddick;Jeng-Shyang Pan

  • Affiliations:
  • Department of Information Management, 840, Chengching Rd., Niausung, Kaohsiung, Taiwan.;School of Computer Science, Engineering and Mathematics, Flinders University, P.O. Box 2100, Adelaide, South Australia 5001, Australia.;Department of Electronic Engineering, National Kaohsiung University of Applied Sciences, Kaohsiung 807, Taiwan

  • Venue:
  • International Journal of Business Intelligence and Data Mining
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper two categories of improvements are suggested thatcan be applied to most k-medoids-based algorithms -conceptual/algorithmic improvements, and implementationalimprovements. These include the revisiting of the accepted casesfor swap comparison and the application of partial distancesearching and previous medoid indexing to clustering. Varioushybrids are then applied to a number of k-medoids-based algorithmsand the method is shown to be generally applicable. Experimentalresults on both artificial and real datasets demonstrate that whenapplied to CLARANS the number of distance calculations can bereduced by up to 98%.