GRAPHICAL METHODS FOR DETERMINING THE PRESENCE OF PERIODIC CORRELATION
Authors:
- 1 July 1991
- journal article
- Published by Wiley in Journal of Time Series Analysis
- Vol. 12 (4) , 337-350
Abstract
Methods for determining whether an observed time series is, or contains, a periodically correlated sequence are presented. They are based on the fact that the support of the spectral measure for a harmonizable periodically correlated stochastic sequence is contained in a set of equally spaced lines parallel to the main diagonal in the bifrequency plane. We show how a coherence statistic devised by N. R. Goodman can be used to test for the presence of spectral support (correlation) along diagonal lines, thus providing a test for the presence of periodic correlation. Plots of this coherence statistic as a function of two frequency variables visually reveal the presence of diagonal support lines. Two additional tests for the presence of diagonal support lines are constructed by collapsing the two‐dimensional coherence plots into one‐dimensional plots of difference frequency.Keywords
This publication has 8 references indexed in Scilit:
- Exploitation of spectral redundancy in cyclostationary signalsIEEE Signal Processing Magazine, 1991
- Representation of strongly harmonizable periodically correlated processes and their covariancesJournal of Multivariate Analysis, 1989
- Nonparametric time series analysis for periodically correlated processesIEEE Transactions on Information Theory, 1989
- A LIMITING PROPERTY OF SAMPLE AUTOCOVARIANCES OF PERIODICALLY CORRELATED PROCESSES WITH APPLICATION TO PERIOD DETERMINATIONJournal of Time Series Analysis, 1988
- Correlation Theory of Stationary and Related Random Functions1987
- Time series with periodic structureBiometrika, 1967
- Periodically and Almost-Periodically Correlated Random Processes with a Continuous Time ParameterTheory of Probability & Its Applications, 1963
- Spectral Analysis of Abstract FunctionsTheory of Probability & Its Applications, 1959