67
18

Deconvolution for an atomic distribution

Abstract

Let X1,...,XnX_1,...,X_n be i.i.d. observations, where Xi=Yi+σZiX_i=Y_i+\sigma Z_i and YiY_i and ZiZ_i are independent. Assume that unobservable YY's are distributed as a random variable UV,UV, where UU and VV are independent, UU has a Bernoulli distribution with probability of zero equal to pp and VV has a distribution function FF with density f.f. Furthermore, let the random variables ZiZ_i have the standard normal distribution and let σ>0.\sigma>0. Based on a sample X1,...,Xn,X_1,..., X_n, we consider the problem of estimation of the density ff and the probability p.p. We propose a kernel type deconvolution estimator for ff and derive its asymptotic normality at a fixed point. A consistent estimator for pp is given as well. Our results demonstrate that our estimator behaves very much like the kernel type deconvolution estimator in the classical deconvolution problem.

View on arXiv
Comments on this paper