Wiener series

In mathematics, the Wiener series (or Wiener G-functional expansion) originates from the 1958 book of Norbert Wiener. It is an orthogonal expansion for nonlinear functionals closely related to the Volterra series and having the same relation to it as an orthogonal polynomial expansion has to a power series. The analogue of the coefficients are referred to as Wiener kernels. The terms of the series are orthogonal (uncorrelated) with respect to a statistical input of white noise. This property allows the terms to be identified in applications by the Lee-Schetzen method.

The Wiener series is important in nonlinear system identification. In this context, the series approximates the functional relation of the output to the entire history of system input at any time. The Wiener series has been applied mostly to the identification of biological systems, especially in neuroscience.

The name Wiener series is almost exclusively used in system theory. In the mathematical literature it occurs as the Ito expansion (1951) which is entirely equivalent to it. (The Wiener series should not be confused with the Wiener filter, which is an unrelated concept.)

Wiener G-functional expression

$G_{0}x(n)=k_{0}=E\left\{{y(n)}\right\};$

being $y[n]$ the system output,

$G_{1}x(n)=\sum _{\tau _{1}=0}^{N_{1}-1}{k_{1}(\tau _{1})x(n-\tau _{1})};$

$G_{2}x(n)=\sum _{\tau _{1},\tau _{2}=0}^{N_{2}-1}{k_{2}(\tau _{1},\tau _{2})x(n-\tau _{1})x(n-\tau _{2})}-A\sum _{\tau _{1}=0}^{N_{2}-1}{k_{2}(\tau _{1},\tau _{1})};$

$G_{3}x(n)=\sum \limits _{\tau _{1},\ldots ,\tau _{3}=0}^{N_{3}-1}{k_{3}(\tau _{1},\tau _{2},\tau _{3})x(n-\tau _{1})x(n-\tau _{2})x(n-\tau _{3})}-3A\sum _{\tau _{1}=0}^{N_{3}-1}\sum _{\tau _{2}=0}^{N_{3}-1}k_{3}(\tau _{1},\tau _{2},\tau _{2})x(n-\tau _{1});$

$G_{4}x(n)=\sum \limits _{\tau _{1},\ldots ,\tau _{4}=0}^{N_{4}-1}k_{4}(\tau _{1},\tau _{2},\tau _{3},\tau _{4})x(n-\tau _{1})x(n-\tau _{2})x(n-\tau _{3})x(n-\tau _{4})+$

$-6A\sum \limits _{\tau _{1},\tau _{2}=0}^{N_{4}-1}\sum \limits _{\tau _{3}=0}^{N_{4}-1}{k_{4}(\tau _{1},\tau _{2},\tau _{3},\tau _{3})x(n-\tau _{1})x(n-\tau _{2})}+3A^{2}\sum \limits _{\tau _{1},\tau _{2}=0}^{N_{4}-1}{k_{4}(\tau _{1},\tau _{1},\tau _{2},\tau _{2})};$

$G_{5}x(n)=\sum \limits _{\tau _{1}\ldots \tau _{5}=0}^{N_{5}-1}k_{5}(\tau _{1},\tau _{2},\tau _{3},\tau _{4},\tau _{5})x(n-\tau _{1})x(n-\tau _{2})x(n-\tau _{3})x(n-\tau _{4})x(n-\tau _{5})+$

$-10A\sum \limits _{\tau _{1},\ldots \tau _{3}=0}^{N_{5}-1}\sum \limits _{\tau _{4}=0}^{N_{5}-1}k_{5}(\tau _{1},\tau _{2},\tau _{3},\tau _{4},\tau _{4})(\tau _{1},\tau _{2},\tau _{3})x(n-\tau _{1})x(n-\tau _{2})x(n-\tau _{3})$

$+15A^{2}\sum \limits _{\tau _{1}=0}^{N_{5}-1}\sum \limits _{\tau _{2},\tau _{3}=0}^{N_{5}-1}k_{5}(\tau _{1},\tau _{2},\tau _{2},\tau _{3},\tau _{3})(\tau _{1})x(n-\tau _{1});$

where A is the input variance.

See also

• Volterra series
• System identification
• Spike-triggered average

References

• Wiener, Norbert (1958). Nonlinear Problems in Random Theory. Wiley and MIT Press.
• Lee and Schetzen; Schetzen‡, M. (1965). "Measurement of the Wiener kernels of a non-linear system by cross-correlation". International Journal of Control. First. 2 (3): 237–254. doi:10.1080/00207176508905543.
• Marmarelis, P.Z.; Naka, K. (1972). "White-noise analysis of a neuron chain: an application of the Wiener theory". Science. 175 (4027): 1276–1278. doi:10.1126/science.175.4027.1276. PMID 5061252.
• Schetzen, Martin (1980). The Volterra and Wiener Theories of Nonlinear Systems. John Wiley and Sons. ISBN 978-0-471-04455-0.
• Marmarelis, P.Z. (1991). "Wiener Analysis of Nonlinear Feedback". Sensory Systems Annals of Biomedical Engineering. 19: 345–382.
• Franz, M; Schölkopf, B. (2006). "A unifying view of Wiener and Volterra theory and polynomial kernel regression". Neural Computation. 18 (12): 3097–3118. doi:10.1162/neco.2006.18.12.3097.