Conditional versions of limit theorems for conditionally associated random variables

From MaRDI portal





A concept of conditional association of a sequence of random variables introduced recently by Prakasa Rao is investigated. A sequence \(X_1,\dots, X_n\) of random variables is said to be conditionally associated given sub-\(\sigma\)-algebra \({\mathcal F}\) if \(\text{Cov}^{{\mathcal F}}(f(X_1,\dots, X_n), g(X_1,\dots, X_n))\geq 0\) a.s. for any two coordinate-wise nondecreasing functions \(f\), \(g\) on \(\mathbb{R}^n\), whenever the conditional covariance exists. It is shown that associated random variables need not be conditionally associated and vice versa. Next, several basic properties and theorems which hold for associated random variables are obtained for the case of conditionally associated random variables. Among them, conditional Hájek-Rényi type inequalities are proved and applied to derive a conditional strong law of large numbers. Also, a conditional central limit theorem stated in terms of conditional characteristic functions is established. Some auxiliary lemmas may be of independent interest.



Cites work


Cited in
(25)








This page was built for publication: Conditional versions of limit theorems for conditionally associated random variables

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q624584)