Spring Seminar Series - January 19, 2006
University of Minnesota
School of Statistics
College of Liberal Arts

Estimation of a Semiparametric Transformation Model

Ingrid van Keilegom
Department of Statistics
Universite Catholique de Louvain, Belgium

Thursday, January 19, 2006
3:30 PM, 115 Ford Hall
Minneapolis, East Bank Campus
Social at 3:00 PM, 300 Ford Hall

Abstract

We propose consistent estimators for transformation parameters in semiparametric models. The problem is to find the optimal transformation into the space of models with a predetermined structure. Special cases are searching for the transformation to yield an additive or a multiplicative nonparametric model. We do not focus on a particular non-parametric estimator, but give results for the estimation of the transformation when the rest of the model is estimated non- or semi-parametrically and fulfills some consistency conditions. For the estimation of the transformation parameter are proposed two methods: minimizing the mean squared distance from independence or maximizing a profiled likelihood function. First is discussed the problem of identification of such models. We then state results for a general class of nonparametric estimators. Finally we give some particular examples for nonparametric estimation of transformed separable models. The theoretical results as well as the small sample performance are studied by several simulation exercises.

This work is joint work with Oliver Linton and Stefan Sperlich.