Statistics Seminar: Justin Wishart -- Kink estimation with Long-Range dependent noise

Justin Wishart School of Mathematics and Statistics University of Sydney 

Location: Carslaw 173 

Time: 2pm Friday, March 27, 2009 

Title: Kink estimation with Long-Range dependent noise 

Abstract: In this seminar we study the non-parametric estimation of the location of jump
points in the first derivative (referred to as kinks) of a regression function f in the
presence of noise that exhibits long-range dependence (LRD).  In particular, we consider
the situation when observations are derived from a direct model with regular design
points.  The method is based on the zero-crossing technique and makes use of high-order
kernels and is optimal in the minimax sense.  The kink location and estimation technique
is demonstrated on some simulated data and the detrimental effect of LRD on the rate of
convergence is shown.  We also apply our kink analysis on Australian temperature dataset
as an example.