SADDLEPOINT METHODS FOR BOOTSTRAP CONFIDENCE BANDS IN NONPARAMETRIC REGRESSION
Author:
- 1 March 1993
- journal article
- Published by Wiley in Australian Journal of Statistics
- Vol. 35 (1) , 93-101
Abstract
Summary: Bootstrap techniques have been used to construct confidence bands in nonparametric regression problems (Härdle & Bowman, 1988). Yet the required simulation is generally computationally intensive and therefore makes it difficult to conduct further investigations. In this paper, two saddlepoint methods are considered as alternatives to the naive simulation procedure. Some improvements to Härdle & Bowman's bootstrap method are suggested. The improvements are numerically verified using these efficient and accurate analytic methods.Keywords
This publication has 14 references indexed in Scilit:
- General saddlepoint approximations in the bootstrapStatistics & Probability Letters, 1992
- Bootstrap Simultaneous Error Bars for Nonparametric RegressionThe Annals of Statistics, 1991
- Small-Sample Confidence IntervalsJournal of the American Statistical Association, 1990
- On nonparametric regression with higher-order kernelsJournal of Statistical Planning and Inference, 1989
- Local Bandwidth Selection for Kernel EstimatesJournal of the American Statistical Association, 1989
- Saddlepoint approximations in resampling methodsBiometrika, 1988
- Bootstrapping in Nonparametric Regression: Local Adaptive Smoothing and Confidence BandsJournal of the American Statistical Association, 1988
- Residual variance and residual pattern in nonlinear regressionBiometrika, 1986
- General Saddlepoint Approximations with Applications to L StatisticsJournal of the American Statistical Association, 1986
- Confidence Bands for Regression FunctionsJournal of the American Statistical Association, 1985