HOS or SOS for parametric modeling?
- 1 January 1991
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 15206149,p. 3097-3100 vol.5
- https://doi.org/10.1109/icassp.1991.150110
Abstract
Parametric models obtained via second-order statistics (SOS) are appropriate when the available stationary data are linear, Gaussian, and time-reversible. On the other hand, evidence of nonlinearity, non-Gaussianity, or time-irreversibility favors the use of higher-order statistics (HOS). To quantify normality and time-reversibility, and thus resolve the title question, consistent, time-domain statistical tests are developed and analyzed in a Neyman-Pearson framework. The novel test statistics are computationally attractive and streamlined towards parametric modeling because they employ the minimal HOS lags which uniquely characterize autoregressive moving-average processes. Simulations illustrate the performance of the proposed tests.This publication has 7 references indexed in Scilit:
- Signal detection and classification using matched filtering and higher order statisticsIEEE Transactions on Acoustics, Speech, and Signal Processing, 1990
- On identifiability, maximum-likelihood, and novel HOS based criteriaPublished by Institute of Electrical and Electronics Engineers (IEEE) ,1990
- Asymptotically optimal estimation of MA and ARMA parameters of non-Gaussian processes from high-order momentsIEEE Transactions on Automatic Control, 1990
- On time-reversibility and the uniqueness of moving average representations for non-Gaussian stationary time seriesBiometrika, 1988
- TESTING FOR GAUSSIANITY AND LINEARITY OF A STATIONARY TIME SERIESJournal of Time Series Analysis, 1982
- A TEST FOR LINEARITY OF STATIONARY TIME SERIESJournal of Time Series Analysis, 1980
- Time-reversibility of linear stochastic processesJournal of Applied Probability, 1975