NON-PARAMETRIC MAXIMUM LIKELIHOOD ESTIMATION
Gordon B. Crawford, Sam C. Saunders · 1963
Abstract : Given that a distribution function is a member of a subclass of absolutely continuous measures, the problem of nonparametric estimation is considered, with the method of maximum likelihood, of the underlying density function of a given sample of independent identically distributed random variables. Sufficient conditions on the space of probability densities and its topology are given for the consistency of such an estimate.