Μη παραμετρική συμπερασματολογία για σταθμισμένες κατανομές
View/ Open
Subject
Σύνθετες κατανομές ; Συναρτήσεις ; ΣτατιστικήAbstract
The present thesis consists of a review of the literature on nonparametric and semi-parametric estimation of an unknown distribution function based on several biased samples. In the first part the nonparametric estimation of a distribution function is studied when the data arise from weighted versions of it with known weighting functions and algorithms for the calculation of its nonparametric max¬imum likelihood estimators are presented. In the second part the case where the underlying distributions satisfy the so-called density ratio model is considered and the semi-parametric estimation of the corresponding distribution functions is inves¬tigated as well as tests of hypotheses about the equality of these distributions are presented.