Weighted dissimilarity measures for binary data

Research output: Article

Abstract

The weighted dissilimiarity index (WDI), which defines the 'importance' of attributes as a linear function of their presence, is generalized. A new weight function based on Shannon's entropy is proposed, and the notion of relative and absolute weights is introduced. The relationship of weighted dissilimarities to classical binary dissilimarities is discussed. Differences between the uniform, linear and the proposed nonlinear weight are demonstrated by an artificial example which represents a continuous change of species composition. As a field example, the zones of a beechwood vegetation are separated along a 63 m long transect using multidimensional scaling.

Original languageEnglish
Pages (from-to)105-108
Number of pages4
JournalAbstracta Botanica
Volume20
Issue number2
Publication statusPublished - dec. 1 1996

ASJC Scopus subject areas

  • Agricultural and Biological Sciences(all)

Fingerprint Dive into the research topics of 'Weighted dissimilarity measures for binary data'. Together they form a unique fingerprint.

  • Cite this