Title: Density estimators that can be plugged in
Authors: Eric Beutner - Vrije Universiteit Amsterdam (Netherlands) [presenting]
Henryk Zaehle - Saarland University (Germany)
Abstract: Density estimators that can be plugged in are considered. The notion plug-in property of density estimators has been introduced in 2003. It means that statistical functionals like the expected value or the second moment can be estimated by replacing the density by its estimate in the respective formula. We focus on kernel based methods and show that appropriate kernels have the plug-in property for large classes of functionals. It will be demonstrated that the results continue to hold for many stationary time series models.