The theorem on deviation of empirical measure and its applications

In the paper we obtain the moderate deviation theorems for the empirical measures in one- and multi-dimensional cases. The results are used in the construction of the density estimates having the best rates of convergence in a uniform metric.