Estimates of the rate of convergence in the limit theorem for negative binomial random sums (Q6159084): Difference between revisions

Let \(N(t)\) have the negative binomial distribution with parameters \(r,t>0\) and mass function \[ P(N(t)=k)=\frac{\Gamma(r+k)}{k!\Gamma(r)(1+t)^r}\left(\frac{t}{1+t}\right)^k\,, \] for \(k=0,1,\ldots\). Let \(X_1,X_2,\ldots\) be a sequence of independent and identically distributed random variables, independent of \(N(t)\), with zero mean and finite \((2+\delta)\)th moment for some \(0<\delta\leq1\). The author considers the negative binomial random sum defined by \[ S(t)=\sum_{i=1}^{N(t)}X_i\,, \] establishing rates of convergence of \(S(t)\) to a scale mixture of Gaussian laws in the uniform metric. The cases \(r<\delta/2\) and \(r=\delta/2\) are considered separately. In the former case an upper bound of order \(O(t^{-r})\) is given, while in the latter case it is of order \(O(t^{-\delta/2}\log(t))\) as \(t\to\infty\).