"Density estimation and local dependence structures"

			(Martin Sk÷ld)

Abstract:
Continuous-time modelling provides a way of describing how local
dependence structures in data affects the performance of the kernel
density estimator. We will review some results from continuous-time
estimation, relating sample path smoothness to asymptotic variance, and
discuss applications to discretely sampled data and to time-series
observed with a measurement error.