Abstract
A continuous stationary signal possessing non-Gaussian higher order statistics cannot be correctly modelled by any discrete process based on passing independently and identically distributed noise through a linear filter. In particular, it is shown that at third order there exists no discrete skewed linear model with a discrete bispectrum that is the same as that obtained from the Nyquist samples of any continuous stationary process. The nature of the problem is elucidated and an alternative method for modelling the third order statistics of continuous stationary processes is proposed.
Similar content being viewed by others
References
T. Subba Rao and M. Gabr (1980), A test for linearity of stationary time series,J. Time Series Anal. 1, 145–148.
M. J. Hinich (1982), Testing for Gaussianity and linearity of a stationary time series,J. Time Series Anal. 3(3), 169–176.
D. Brillinger and M. Rosenblatt (1967a,b), Asymptotic theory ofkth order spectra, Spectral Analysis of Time Series (ed. B. Harris), 153–188, Wiley. Also; Computation and interpretation ofkth order spectra, 189–232.
J. W. Van Ness (1965), Asymptotic normality of bispectral estimates,Ann. Math. Stat. 37, 1257–1275.
M. J. Hinich and M. A. Wolinsky (1988), A test for aliasing using bispectral components,J. Amer. Stat. Assoc. Theory and Methods 83(402), 499–502.
M. J. Hinich, D. Marandino, and E.J. Sullivan (1989), Bispectrum of ship radiated noise,J. Acoust. Soc. Amer. 85(4), 1512–1517.
M. J. Hinich (1991), Private communication.
Author information
Authors and Affiliations
Additional information
Supported by Defense Research Agency, Portland.
Rights and permissions
About this article
Cite this article
Parsons, A.T., Williams, M.L. Limitations on the use of discrete linear models of continuous random processes. Circuits Systems and Signal Process 13, 403–410 (1994). https://doi.org/10.1007/BF01183738
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF01183738