Abstract:
The problem of nonparametric estimation of a probability distribution is considered for a case when the sample is contaminated by a random noise. It is supposed that the distribution of the error is known only partially. Identifiability and consistent estimation are investigated.