Advertisement

A Voronoi diagram based adaptive k-means-type clustering algorithm for multidimensional weighted data

  • Thomas Schreiber
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 553)

Abstract

This paper describes a solution to the following problem: Given a set of weighted data points, find the cluster center points, which minimize the least squared errors. The k-means-type methods produce good results, but usually the quality of the representation depends on an initial cluster configuration. Also this does not allow a variable number of clusters for a given error tolerance.

The proposed method removes these disadvantages by an adaptive sequential insertion of new clusters in those areas, where the largest errors occur. This can be done more efficiently by using multidimensional Voronoi diagrams and local procedures. The data points can be weighted and arbitrarily distributed in the Euclidean space. The weight of each point may be chosen by the user depending on the importance or correctness of that point. At the same time the method produces a hierarchical multidimensional triangulation of the data at different levels of accuracy.

Keywords

Clustering algorithm Delaunay triangulation hierarchical representation k-means algorithm Voronoi diagram 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    A. Bowyer: Computing Dirichlet tesselations, Comp. Journal, Vol. 24, No. 2, 1981, 162–166.Google Scholar
  2. [2]
    T. Dierks: The Modelling and Visualization of Scattered Volumetric Data, Masters Thesis, Arizona State University, USA, Dec. 1990.Google Scholar
  3. [3]
    J. A. Hartigan: Clustering Algorithms, Wiley, New York, 1975.Google Scholar
  4. [4]
    P. Heckbert: Color image quantization for frame buffer display, ACM Trans. Computer Graphics 16, 3 (July 1982), 297–304.Google Scholar
  5. [5]
    I. Hyafil, R. L. Rivest: Construction optimal binary decision trees is NP-complete, Inf. Process. Lett. 5, May 1976, 15–17.Google Scholar
  6. [6]
    J. R. McMahon: Knot Selection for Least Squares Approximation using Thin Plate Splines, Masters Thesis, Naval Postgraduate School, Monterey, USA, June 1986.Google Scholar
  7. [7]
    J. R. McMahon, R. Franke: An Enhanced Knot Selection Algorithm for Least Squares Approximation using Thin Plate Splines, ARO Report 90-1, Trans. of the Seventh Army, Conf. on Applied Math. and Computing.Google Scholar
  8. [8]
    O. Palacios-Velez, B. C. Renaud: A Dynamic Hierarchical Subdivision Algorithm for Computing Delaunay Triangulations and Other Closest-Point Problems, ACM Trans. on Math. Soft., Vol. 16, No. 3, Sept. 1990, 275–292.Google Scholar
  9. [9]
    F. P. Preparata, M. I. Shamos: Computational Geometry, Springer, 1985.Google Scholar
  10. [10]
    S. Z. Selim, M. A. Ismail: K-means-type algorithms: A generalized convergence theorem and characterization of local optimality, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-6, 1 (1986), 81–87.Google Scholar
  11. [11]
    S. J. Wan, S. K. M. Wong, P. Prusinkiewicz: An Algorithm for Multidimensional Data Clustering, ACM Trans. on Math. Soft., Vol. 14, No. 2, June 1988, 153–162.Google Scholar
  12. [12]
    D. F. Watson: Computing the n-dimensional Delaunay tesselation with application to Voronoi polytops, Comp. Journal, Vol. 24, No. 2, 1981, 162–166.Google Scholar
  13. [13]
    X. Wu, I. H. Witten: A fast k-means type clustering algorithm, Dept. Computer Science, Univ. of Calgary, Canada, May 1985.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1991

Authors and Affiliations

  • Thomas Schreiber
    • 1
  1. 1.Fachbereich InformatikUniversität KaiserslauternKaiserslautern

Personalised recommendations