Description
| Abstract: | This paper is concerned with nonlinear transfer function analysis. In particular, two test statistics are proposed to detect nonlinearity of a transfer function; the first test is of likelihood-ratio type and the second test uses bispectrum and cross-bispectrum. A simulation study shows that the tests work reasonably well in finite samples for certain nonlinear models. We then discuss modeling nonlinear transfer functions and the contribution of nonlinear components in a given model. Several examples are used to illustrate the tests proposed and an example is given to show that substantial gain may result from using simple nonlinear transfer functions. |
|---|
| Item Description: | "August, 1991."
Funding information taken from leaf i. |
|---|
| Physical Description: | 19 leaves, 4 unnumbered leaves : illustrations ; 28 cm |
|---|
| Bibliography: | Includes bibliographical references (leaves 16-17). |
|---|