Improving the accuracy of estimation of unknown random variable probability density over empirical data

A problem under study is to estimate an unknown probability density of a ra ndom variable basing on results of observations. Parzen-Rozenblatt estimate s have been selected to solve the problem. A method improving the accuracy of the empirical density is suggested. It is based on extrapolation over a number of independent observations. Effectiveness of the method is demonstr ated on an example of its application in a model problem. The method develo ped can be applied to estimate spectra of secondary particles while modelin g processes of interactions of high energy particles in matter, in particul ar, while defining characteristics of neutron fields in electronuclear inst allations under study. (C) 2001 Elsevier Science B.V. All rights reserved.