Möbus, Claus (1981) On the Nontermination of the K-Means Clustering Algorithm for Certain Data Sets. EDV in Medizin und Biologie, 12 (4). pp. 124-126. ISSN 0300-8282
|
- Published Version
Volltext (1174Kb) |
Abstract
There are some well accepted standards for the formulation of algorithms and computer programs (Bauer and Wössner, 1981): (a) efficiency, (b) finiteness of description and (c) finiteness of operation. It can be demonstrated that the popular K-Means clustering algorithm loops infinitely for some data sets if the computer program is only a result of a mechanical translation of its mathematical correct form. The resulting program contradicts postulate (c) because of the finite precision of computer arithmetic. Besides its theoretical aspects the article has a practical purpose. It shows ways of correcting some published FORTRAN-programs so that termination is achieved for all data sets.
Item Type: | Article |
---|---|
Uncontrolled Keywords: | K-Means clustering algorithm, finiteness of operation, finite precision, nontermination |
Subjects: | Generalities, computers, information > Computer science, internet Philosophy and psychology > Psychology |
Divisions: | School of Computing Science, Business Administration, Economics and Law > Department of Computing Science |
Date Deposited: | 23 Oct 2015 09:10 |
Last Modified: | 23 Oct 2015 09:10 |
URI: | https://oops.uni-oldenburg.de/id/eprint/2532 |
URN: | urn:nbn:de:gbv:715-oops-26134 |
DOI: | |
Nutzungslizenz: |
Actions (login required)
View Item |