On the Ghurye-Olkin-Zinger theorem (Q2253980)

Characterization theorems in probability theory and mathematical statistics are theorems that establish a connection between the type of the distribution of random variables or random vectors and certain general properties of functions on them. The assumption that two linear (or non-linear) statistics are independent (or identically distributed, or have a constant regression, and so on) can be used to characterize various populations. For example, according to S. N. Bernstein characterization theorem, if \(X_1\) and \(X_2\) are independent random variables and the statistics \(S_1 = X_1 + X_2\) and \(S_2 = X_1 -X_2\) are independent, then the variables \(X_1\) and \(X_2\) have a normal distribution. The next step in this direction is the Skitovich-Darmois theorem, in which are analyzed \(n\) independent random variables and two linear statistics \(S_1\) and \(S_2\). \textit{Yu. V. Linnik} [Decompositions of probability laws (Russian). Leningrad: Izdat (1960; Zbl 0093.15002)], and \textit{B. Ramachandran} [Advanced theory of characteristic functions (Russian). Moskau: Verlag `Nauka', Hauptredaktion für physikalisch-mathematische Literatur (1975; Zbl 0303.60013)] extended the Skitovich-Darmois theorem. Another generalization of the Skitovich-Darmois theorem belongs to \textit{S. G. Ghurye} and \textit{I. Olkin} [Ann. Math. Stat. 33, 533--541 (1962; Zbl 0109.13401)]. Finally, a generalization of the Ghurye-Olkin-Zinger characterization theorem is analyzed in this paper.