Stein factors for negative binomial approximation in Wasserstein distance (Q2348733)

Essential ingredients of Stein's method for probability approximation include bounds on the solution of the Stein equation (referred to as Stein factors, or magic factors). This paper gives such bounds for negative binomial approximation in the Wasserstein distance. The proof of the Stein factors is probabilistic, and builds upon techniques developed in [\textit{A. D. Barbour} and \textit{A.-H. Xia}, Bernoulli 12, No. 6, 943--954 (2006; Zbl 1328.62076)] to find analogous bounds in the Poisson case. These Stein factors are applied to give an explicit bound in the negative binomial approximation of the number of parasites in a host under a model which allows for a varying rate of ingestion of parasites over time. The Wasserstein metric is useful here since the infectivity of a host can be expected to be proportional to the number of parasites it carries.