Nonparametric Spectrum Estimation for Spatial Data
Published February 2006
Smoothed nonparametric kernel spectral density estimates are considered for stationary data observed on a d-dimensional lattice. The implications for edge effect bias of the choice of kernel and bandwidth are considered. Under some circumstances the bias can be dominated by the edge effect. We show that this problem can be mitigated by tapering. Some extensions and related issues are discussed. MSC: 62M30, 62M15 C22
Paper Number EM/2006/498:
Download PDF - Nonparametric Spectrum Estimation for Spatial Data
JEL Classification: C22