Statistical analysis of some random truncation models

Abstract

Statistical analysis of some method with empirical Statistical information collected from an experimental distribution of a data set and in many fields such as statistical mechanics and economics has attracted the attention of many authors. The empirical (E), the unknown probabilities it estimates, are always negative. Unknown parameters estimated from this method have a bounded influence function. In this paper, we investigate the empirical (E) method in the truncated data and we prove that it has an asymptotic distribution, a weighted chi-square. We show that when the data is contaminated, empirical (E) method behaves better than empirical likelihood (EL) method and has a better coverage probability