A Note on the Inverse Moment for the Non Negative Random Variables

Shuhe Hu,Xinghui Wang,Wenzhi Yang,Xuejun Wang
DOI: https://doi.org/10.1080/03610926.2012.673677
2014-01-01
Abstract:Let {Zn} be a sequence of non negative random variables satisfying a Rosenthal-type inequality and , where {Mn} is a sequence of positive real numbers. By using the Rosenthal-type inequality, the inverse moment E(a + Xn)− α can be asymptotically approximated by (a + EXn)− α for all a > 0 and α > 0. Furthermore, we show that E[f(Xn)]− 1 can be asymptotically approximated by [f(EXn)]− 1 for a function f( · ) satisfying certain conditions. Our results generalize and improve some corresponding results, which can allow immediate applications to compute the inverse moments for the non negative random variables whose distributions are such as Binomial distribution, Poisson distribution, Gamma distribution, etc.
What problem does this paper attempt to address?