This paper is concerned with the estimation of a varying-coefficient partially linear regression model that is frequently used in statistical modeling. We first construct estimators of the parametric components and the error variance by a wavelet procedure and establish their asymptotic normalities under weaker conditions than those assumed in the previous literature. Then we propose appropriate estimators for the functions characterizing the nonlinear part of the model and derive their convergence rates. Furthermore, we present consistent estimators for the asymptotic (co)variances of the parametric components and error variance estimators as well. These results can be used to make asymptotically valid statistical inference.
|Number of pages||14|
|Journal||Statistics and Probability Letters|
|Publication status||Published - 1 Jun 2004|
- Asymptotic normality
- Least-squares estimation
- Partially linear regression model