SMOOTHED BOOTSTRAP BANDWIDTH SELECTION IN NONPARAMETRIC DENSITY ESTIMATION FOR MOVING AVERAGE PROCESSES

Abstract

This article is concerned with the problem of selecting a suitable bandwidth when estimating the marginal density function of a moving average process. The smoothed bootstrap method is used to implement a bandwidth selector for a convolution-type density estimator, based on the kernel method. The relative rate of convergence of this selector with respect to the MISE bandwidth is proved to be O _P (n ^−1/2). The finite sample size performance of the selector is investigated in a simulation study.

Acknowledgments