Asymptotic results for FGM random sequences (Q5953885): Difference between revisions

Consider a sequence of random variables \((X_n)\) and denote by \(\overline{X}_n=(X_1+X_2+\dots +X_n)/n\) and by \(M_n=\max_{1 \leq j \leq n}\{X_j\} .\) Here, the random variables follow a special dependence structure of the so-called Farlie-Gumbel-Morgenstern typ [see e.g. \textit{S. Cambanis}, in: Advances in probability distributions with given marginals. Math. Appl. 67, 207-222 (1991; Zbl 0744.60037)]. Under this dependence structure and additional assumptions a strong law of large numbers for \(\overline{X}_n \) and an almost sure version of the weak limit relation \(a_n M_n+b_n@>d >>H\) with a max-stable distribution \(H\) and normalizing sequences \((a_n)>0\) and \((b_n)\) are shown.