Gaussian approximation theorems for urn models and their applications (Q1872367)

The authors consider weak and strong Gaussian approximations for a two-color generalized Friedman's urn model with homogeneous and nonhomogeneous generating matrices. The authors present functional central limit theorems and the laws of iterated logarithm. As an application, asymptotic properties for the randomized play-the-winner rule are obtained. Based on the Gaussian approximations, the authors also get some variance estimators for the urn model.