On a nonparametric test for linear relationships
β Scribed by Holger Dette
- Publisher
- Elsevier Science
- Year
- 2000
- Tongue
- English
- Weight
- 108 KB
- Volume
- 46
- Category
- Article
- ISSN
- 0167-7152
No coin nor oath required. For personal study only.
β¦ Synopsis
In a recent paper Azzalini and Bowman (1993, J. Roy. Statist. Soc. Ser. B 55, 549 -559) proposed a pseudolikelihood ratio test for checking the linearity in a homoscedastic nonparametric regression model under a ΓΏxed design assumption. In this paper, we study the asymptotic properties of this test and establish asymptotic normality under the hypothesis of linearity and under any ΓΏxed alternative with di erent rates of convergence in both cases. In a second part of the paper these results are extended to nonparmetric regression models with a random design.
π SIMILAR VOLUMES
In this paper the interest is in testing whether a regression function is a polynomial of a certain degree. One possible approach to this testing problem is to do a parametric polynomial ΓΏt and a nonparametric ΓΏt and to reject the null hypothesis of a polynomial function if the distance between the
We study the intervals of linearity for the regression function of stationary and strongly mixing vectorial-valued random processes, using nonlinear integral plug-in estimators of functionals of the second derivative of the regression function. We give conditions in order to obtain CLT, possibly deg