RATES OF CONVERGENCE IN SEMI‐PARAMETRIC MODELLING OF LONGITUDINAL DATA

• DOI: 10.1111/j.1467-842x.1994.tb00640.x
• Publication Date: 2008-02-26T10:56:19Z
• Authors: R.A. MOYEED; P.J. DIGGLE
• Keywords: Regression function; Kernel regression; Kernel (algebra); Semiparametric regression

Summary We consider the problem of semi‐parametric regression modelling when data consist a collection short time series for which measurements within are correlated. The objective is to estimate function form E[Y( t ) | x ] = x'ß+ μ( ), where μ(.) an arbitrary, smooth , and vector explanatory variables may or not vary with t. For non‐parametric part estimation we use kernel estimator fixed bandwidth h. When h chosen without reference give exact expressions bias variance estimators β μ(t) asymptotic analysis case in number tends infinity whilst per held fixed. also report results small‐scale simulation study indicate extent theoretical continue hold by data‐based cross‐validation method.