Asymptotic efficient estimation in semiparametric nonlinear regression models

In this paper, the estimation method based on the “generalized profile likelihood” for the conditionally parametric models in the paper given by Severini and Wong (1992) is extended to fixed design semiparametric nonlinear regression models. For these semiparametric nonlinear regression models,the resulting estimator of parametric component of the model is shown to be asymptotically efficient and the strong convergence rate of nonparametric component is investigated. Many results (for example Chen (1988), Gao & Zhao (1993), Rice (1986) et al.) are extended to fixed design semiparametric nonlinear regression models. © 1999, Springer Verlag. All rights reserved.