Convergence rates of kernel density estimates in particle filtering

Abstrakt

Bounds on convergence rates of kernel density estimates in particle filtering are specified. The kernel density estimates are shown to be efficient for the Sobolev class of filtering densities.

The upper bounds are established using Fourier analysis whilst the lower ones rely on tools of information theory. (C) 2019 Elsevier B.V. All rights reserved.

Klíčová slova

• Particle filtering
• Kernel density estimates
• Convergence rates