Abstract:
This study uses Steins method and w -functions to determine new bounds, nonuniform bounds, for approximating the distribution and cumulative distribution function of nonnegative integer-valued random variable by the distribution and cumulative distribution function of geometric random variable. For applications, this study uses the obtained results to approximate hypergeometric, Pólya and negative Pólya distributions. By comparing the new bounds of this study and the old bounds in Teerapabolarn (2011), it is found that the new bounds are better than the old bounds, that is, the new bounds can be measured the accuracy of the approximation to be better than the old bounds both theoretical and applications. In addition, measuring the accuracy of the approximation of the cumulative distribution function of nonnegative integer-valued random variable by the cumulative distribution function of geometric random variable, the new bounds have no the restriction of the value of q that are in 1 2 (0, 1/2] or 2 3 (0, 2/3] .