Nearest-neighbor analysis of a family of fractal distributions (Q751698)

In this study of the nearest-neighbour behavior of a family of distributions on the unit interval [0,1), Ibragimov's celebrated central limit theorem for entropy has been exploited to obtain limit theorems for linear normalizations of the log minimum distance when the observations come from fractal distributions. Under the strong mixing condition, it has been shown that in ``almost all'' cases a particular linear transformation of the nearest-neighbour distance converges in distribution to a Gaussian distribution with variance that depends on the entropy and Hausdorff dimension of the original measure. Bounds on the asymptotic distribution of the log minimum distance have been given in some special cases where this asymptotic normality fails.