A limit theorem for the maximum of random variables with logarithmic distribution
Keywords:
Extremes, Limit theorem, Linear normalization, Logarithmic distributionAbstract
Let $X_1, X_2, \ldots, X_n$ be independent and identically distributed random variables and let $M_n = \max \left(X_1, \ldots, X_n\right)$ denote their maximum. In the case of discrete distribution of $X_i$ an appropriate normalization for a non-degenerate limiting distribution of $M_n$ does not always exists. In such a case is usually used the scheme of series. In this note, we derive a non-degenerate limiting distribution for the maximum of random variables with logarithmic distribution in the scheme of series.
References
C. W. Anderson. Extreme value theory for a class of discrete distributions with applications to some stochastic processes. Journal of Applied Probability, 7, (1970) 99–113.
C. W. Anderson, S. G. Coles, J. Husler. ¨ Maxima of Poisson-like variables and related triangular arrays. Annals of Applied Probability, 7 (1997), 953–971.
M. R. Leadbetter, G. Lindgren, H. Rootzen. ´ Extremes and Related Properties of Random Sequences and Processes, New York, Springer Verlag, 1987.
S. Nadarajah, K. V. Mitov. Asymptotics of maxima of discrete random variables. Extremes, 5 (2002), 287–294.
S. Nadarajah, K. V. Mitov. Extremal limit laws for discrete random variables. Journal of Mathematical Sciences, 122 (2004), 3404–3415.