A generalized real-valued measure of the inequality associated with a fuzzy random variable (Q5928966): Difference between revisions

Fuzzy random variables (FRVs) have been introduced by \textit{M. L. Puri} and \textit{D. A. Ralescu} [J. Math. Anal. Appl. 114, 409-422 (1986; Zbl 0592.60004)] as an extension of random sets, in order to model fuzzy set-valued measurable mappings that convert experimental outcomes into fuzzy sets on certain (Euclidean) spaces. The main purpose of this paper is to introduce a real-valued generalized measure of the ``relative variation'' (or inequality) associated with an FRV, and to investigate its relevant properties. Section 2 exposes preliminary concepts. Section 3 defines this real-valued generalized measure of an FRV, based on Csiczár's \(f\)-divergence notion. The proposed definition encloses as a particular case the extension to FRVs of the additively decomposable indices. Section 4 presents the characterization of the functions on which are based the most operational and valuable inequality indices. Because in the definition of the \(f\)-inequality indices the function \(f\) is assumed to satisfy certain conditions, two main families of inequality measures are examined: (i) the twice-differentiable functions \(f\) that satisfy the equation \(f(u)+ f(1/u)= 0\) for all \(u\in (0,\infty)\); (ii) the twice-differentiable functions that satisfy the inequality \(f(u)+ f(1/u)\geq 0\) for all \(u\in (0,\infty)\), with equality if and only if \(u= 0\). Section 5 of the paper proves some fundamental properties of the proposed generalized measure, and considers separately an outstanding measure for each of the above-mentioned function families (the hyperbolic index and the idex of the Shannon type) for some concrete properties. Section 6 illustrates the utility of the introduced measures on two examples involving the computation and use of certain \(f\)-inequality indices in comparing populations.