Parameter and Functional Estimation in Semiparametric Modelsh

Abstract

This talk contains three parts: In part one, we illustrate a semiparametric approach in estimating parameters in variance function. The method is rooted in a projection operation and possesses a robust feature. An illustrational example is the Leukemia study data. In part two, we demonstrate the properties of a semiparametric estimator in estimating a functional. The method is illustrated in an "Eating at the America's Tables" study. The difference and link of the underlying geometric structure is illustrated in part three. 　　　The talk is partially based on joint works with Ray Carroll, Arnab Maity and Yuedong Wang.