NP-Optimal Kernels for Nonparametric Sequential Detection Rules
Loading...
Date
2004
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universitätsbibliothek Dortmund
Abstract
An attractive nonparametric method to detect change-points sequentially is
to apply control charts based on kernel smoothers. Recently, the strong convergence of the
associated normed delay associated with such a sequential stopping rule has been studied
under sequences of out-of-control models. Kernel smoothers employ a kernel function to
downweight past data. Since kernel functions with values in the unit interval are sufficient
for that task, we study the problem to optimize the asymptotic normed delay over a class
of kernels ensuring that restriction and certain additional moment constraints. We apply
the key theorem to discuss several important examples where explicit solutions exist to
illustrate that the results are applicable.
Description
Table of contents
Keywords
control charts, financial data, nonparametric regression, quality control, statistical genetics