No CrossRef data available.
Published online by Cambridge University Press: 14 November 2011
We obtain inequalities for where Wn = anlX1 + … + annXn, the Xr being independent random variables and the Mn being certain truncated means. We then use these inequalities to study the rate at which this probability tends to zero as N→ ∞, noting that in the special case Wn = (X1 + … + Xn)/n, we obtain the estimate given by L. E. Baum and M. Katz which they show is, in a sense, best possible.
A desire to find an inequality which would lead to the result of Baum and Katz was, indeed, the impetus behind this paper.