Abstract
Semiparametric time series regression is often used without checking its suitability,resulting in an unnecessarily complicated model. In practice, one may encounter computationaldifficulties caused by the curse of dimensionality.The paper suggests that to provide more precisepredictions we need to choose the most significant regressors for both the parametric andthe nonparametric time series components.We develop a novel cross-validation-based modelselection procedure for the simultaneous choice of both the parametric and the nonparametrictime series components, and we establish some asymptotic properties of the model selectionprocedure proposed. In addition, we demonstrate how to implement it by using both simulatedand real examples. Our empirical studies show that the procedure works well.
Original language | English |
---|---|
Pages (from-to) | 321-336 |
Journal | Journal of the Royal Statistical Society Series B |
Volume | 66 |
Issue number | 2 |
DOIs | |
Publication status | Published - 2004 |