Abstract
A test is proposed for assessing the lack of fit of heteroscedastic nonlinear regression models that is based on comparison of nonparametric kernel and parametric fits. A data-driven method is proposed for bandwidth selection using the asymptotically optimal bandwidth of the parametric null model which leads to a test that has a limiting normal distribution under the null hypothesis and is consistent against any fixed alternative. The resulting test is applied to the problem of testing the lack of fit of a generalized linear model.
Original language | English (US) |
---|---|
Pages (from-to) | 485-496 |
Number of pages | 12 |
Journal | Canadian Journal of Statistics |
Volume | 27 |
Issue number | 3 |
State | Published - Sep 1999 |
Externally published | Yes |
Keywords
- Bandwidth selection
- Fit comparison test
- Kernel smoother
- Quasi-likelihood estimator
ASJC Scopus subject areas
- Statistics and Probability