Department of Pattern Recognition

(eng): Considering the framework of weak concordance measures introduced by Liebscher in 2014, we propose and study convex weak concordance measures. This class of dependence measures contains as a proper subclass the class of all convex concordance measures, introduced and studied by Mesiar et al. in 2022, and thus it also covers the well-known concordance measures as Spearman's ρ, Gini's γ and Blomqvist's β. The class of all convex weak concordance measures also contains, for example, Spearman's footrule ϕ, which is not a concordance measure. In this paper, we first introduce basic convex weak concordance measures built in general by means of a single point (u,v)∈▽={(u,v)∈]0,1[2|u≥v} and its transpose (v,u) only. Then, based on basic convex weak concordance measures and probability measures on the Borel subsets of ▽, two rather general constructions of convex weak concordance measures are proposed, discussed and exemplified. Inspired by Edwards et al., probability measures-based constructions are generalized to Borel measures on B(]0,1[2)-based constructions also allowing some infinite measures to be considered. Finally, it is shown that the presented constructions also cover the mentioned standard (convex weak) concordance measures ρ, γ, β, ϕ and provide alternative formulas for them.