Saddlepoint approximation for the kernel regression estimator
Assuming real-valued independent and identically distributed observations, we obtain a classical pointwise saddlepoint approximation for the tail probability of the Nadaraya-Watson regression estimator. This saddlepoint approximation is similar to the one which was first obtained by Daniels (Citation1987) for the sample mean via the method of indirect Edgeworth expansion. A simulation study illustrates the accuracy of the proposed approximation and compares it with the usual Gaussian approximation.
Authors
- C. Joutard
Institutions
- Centre National de la Recherche Scientifique (FR)
- Université de Montpellier (FR)
- Université Paul-Valéry Montpellier (FR)
- Institut Montpelliérain Alexander Grothendieck (FR)
Publication Details
- Journal
- Communication in Statistics- Theory and Methods
- Published
- 2026-11-16
- DOI
- https://doi.org/10.1080/03610926.2026.2756559
- Primary Topic
- Probability and Statistical Research
- Type
- article
- Field-Weighted Citation Impact
- 0.00