On possibilistic version of distance covariance and correlation

István Harmati, R. Fullér

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Distance correlation is a relatively new measure of dependence in probability theory and statistics, which has the great advantage that it gives zero if and only if the variables are independent. In this paper we define its possibilistic version. Namely, we equip each γ -level set of the joint possibility distribution with a uniform probability distribution, then we determine the probabilistic distance covariance and correlation between the marginal distributions. Finally, the possibilistic distance covariance and correlation is computed as the weighted average of these probabilistic measures of dependence.

Original languageEnglish
Title of host publicationStudies in Computational Intelligence
PublisherSpringer Verlag
Pages175-181
Number of pages7
DOIs
Publication statusPublished - Jan 1 2019

Publication series

NameStudies in Computational Intelligence
Volume796
ISSN (Print)1860-949X

    Fingerprint

Keywords

  • Dependence
  • Distance correlation
  • Distance covariance
  • Fuzzy numbers
  • Possibilistic correlation

ASJC Scopus subject areas

  • Artificial Intelligence

Cite this

Harmati, I., & Fullér, R. (2019). On possibilistic version of distance covariance and correlation. In Studies in Computational Intelligence (pp. 175-181). (Studies in Computational Intelligence; Vol. 796). Springer Verlag. https://doi.org/10.1007/978-3-030-00485-9_20