Abstract
It should be noted that for linear bivariate models, the third order cross moments γ
xy (k, 1)= 0 for all k and 1. Hence looking at a table of the third order cross moments one can
easily distinguish bivariate bilinear models from linear bivariate models (assuming the
distribution of noise is Gaussian and series xt had been prewhitened). Again, one can also
distinguish between Tables 2, 3, and 4, ie, between diagonal, subdiagonal, and
superdiagonal bivariate bilinear models by looking at these tables. It may also be possible to
distinguish the lags of the particular model by carefully looking at the table. It may be more
interesting to look at the third order cross moment table for a few more general bivariate
bilinear models.
xy (k, 1)= 0 for all k and 1. Hence looking at a table of the third order cross moments one can
easily distinguish bivariate bilinear models from linear bivariate models (assuming the
distribution of noise is Gaussian and series xt had been prewhitened). Again, one can also
distinguish between Tables 2, 3, and 4, ie, between diagonal, subdiagonal, and
superdiagonal bivariate bilinear models by looking at these tables. It may also be possible to
distinguish the lags of the particular model by carefully looking at the table. It may be more
interesting to look at the third order cross moment table for a few more general bivariate
bilinear models.
| Original language | English |
|---|---|
| Title of host publication | Nonlinear Time Series and Signal Processing |
| Editors | R R Mohler |
| Place of Publication | Berlin |
| Publisher | Springer |
| Pages | 59-74 |
| Number of pages | 15 |
| ISBN (Electronic) | 978-3-540-38837-1 |
| ISBN (Print) | 978-3-540-18861-2 |
| DOIs | |
| Publication status | Published - 1988 |
| Externally published | Yes |
Publication series
| Name | Lecture Notes in Control and Information Sciences |
|---|---|
| Volume | 106 |
| ISSN (Print) | 0170-8643 |