Longitudinal Kernel Regression
Naisyin Wang, (Texas A&M University), email@example.com,
Raymond J. Carroll, (Texas A&M University), firstname.lastname@example.org,
Xihong Lin, (U. Of Michigan), email@example.com, and
Ziding Feng, (Fred Hutchinson Cancer Center), firstname.lastname@example.org
There has been a substantial recent interest in investigating the performance of kernel regression estimator for longitudinal data. Most approaches adopt the strategy of ignoring the within-subject correlation structure. When the cluster sizes remain fixed, a result supporting the use of this ``working independence'' strategy indicates that under the conventional estimation procedure, a correct specification of the correlation structure actually diminishes the asymptotic efficiency. In this presentation, I will discuss an alternative kernel estimating equation that accounts for the within subject correlation. The major gain by the new approach is at variation reduction. For nonparametric curve estimation, the variance of the proposed method is uniformly smaller than that of the most efficient working independence approach. Under the framework of marginal generalized partially linear models, the new estimator is semiparametric efficient in the Gaussian case, and is more efficient than the working independence estimator in non-Gaussian cases.