Polyspectrum modeling using linear or quadratic filters (Q4036244)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 168541
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | Polyspectrum modeling using linear or quadratic filters |
scientific article; zbMATH DE number 168541 |
Statements
Polyspectrum modeling using linear or quadratic filters (English)
0 references
16 May 1993
0 references
Gaussian white noise
0 references
0.7571024894714355
0 references
0.7105038166046143
0 references
0.7044342756271362
0 references
0.7018207907676697
0 references
The modeling of linear and quadratic systems (filters) driven by Gaussian white noise (GWN) and with output having a given polyspectrum (the \(p\)th order spectrum being the Fourier transform of the \(p\)th order cumulant function) is studied. Formulas for the output polyspectrum of quadratic filters driven by GWN are provided. In the linear case the modeling problem is known as factorization problem and it is shown that if a polyspectrum is factorable with non-GWN, it can be also modeled as a quadratic filter driven by GWN; the converse is however false. Furthermore, the output \(p\)th order cumulant function of a finite impulse response (FIR) quadratic filter is of finite extent. On the contrary, if a \(p\)th order cumulant function is of finite extent, then we have a FIR if it is factorable; if there is a quadratic modeling, then there is not necessarily a FIR. Also, contrary to the linear case, two quadratic filters driven by the same input can have the same output polyspectra even if they do not differ only for a time shift.
0 references