CMStatistics 2019: Start Registration
View Submission - CMStatistics
B0821
Title: Quantifying and estimating asymmetric dependence Authors:  Florian Griessenberger - University Salzburg (Austria) [presenting]
Wolfgang Trutschnig - University of Salzburg (Austria)
Abstract: Standard dependence measures considered in the literature like Pearson correlation, Spearman rank correlation or Schweitzer and Wolff's $\sigma$ are symmetric, i.e. they assign each pair of random variables $(X,Y)$ the same dependence as they assign the pair $(Y,X)$. Since dependence structures are in general not symmetric (in contrast to independence, which is a symmetric concept), the classical dependence measures fail to detect asymmetry. The recently developed R-package qad (short for quantification of asymmetric dependence) aims at detecting asymmetries in samples. It estimates the dependence of the second variable on the first one and vice versa, and additionally quantifies the asymmetry of the underlying dependence structure. The main objectives are to sketch the idea underlying the copula-based dependence measure, to present the most relevant mathematical properties of the underlying estimator and to illustrate its capabilities by some examples.