Methods for determining the presence of periodic correlation based on the bootstrap methodology

This paper presents methods for detecting the period of non Gaussian PC processes. A new statistic for testing periodic correlation is proposed. It is based on the bootstrap procedure which is used to estimate confidence intervals of coherence statistic. This method is linked to that of Hurd and Gerr based on Goodman's tests so both methodologies are also compared. It is demonstrated that in some situations the new test appears to be better.

MoreLess

Year of publication:	2003
Authors:	Broszkiewicz-Suwaj, Ewa
Institutions:	Hugo Steinhaus Center for Stochastic Methods, Politechnika Wrocławska
Subject:	Periodic correlation \| Bootstrap \| Spectral representation

More details

Extent:	application/pdf
Series:	HSC Research Reports.
Type of publication:	Book / Working Paper
Language:	English
Notes:	Number HSC/03/02 8 pages
Classification:	C22 - Time-Series Models ; C32 - Time-Series Models
Source:	RePEc - Research Papers in Economics

Persistent link: https://www.econbiz.de/10009003599