Fast Algorithms for Nonparametric Curve Estimation
- 1 June 1994
- journal article
- research article
- Published by Taylor & Francis in Journal of Computational and Graphical Statistics
- Vol. 3 (2) , 192-213
- https://doi.org/10.1080/10618600.1994.10474638
Abstract
Naive implementations of local polynomial fits and kernel estimators require almost O(n 2) operations. In this article two fast O(n) algorithms for nonparametric local polynomial fitting are presented. They are based on updating normal equations. Numerical stability is guaranteed by controlling ill-conditioned situations for small bandwidths and data-tuned restarting of the updating procedure. Restarting at every output point results in a moderately fast but highly stable O(n 7/5) algorithm. Applicability of algorithms is evaluated for estimation of regression curves and their derivatives. The idea is also applied to kernel estimators of regression curves and densities.Keywords
This publication has 12 references indexed in Scilit:
- Local Linear Regression Smoothers and Their Minimax EfficienciesThe Annals of Statistics, 1993
- A Flexible and Fast Method for Automatic SmoothingJournal of the American Statistical Association, 1991
- Discussion: Linear Smoothers and Additive ModelsThe Annals of Statistics, 1989
- Averaged Shifted Histograms: Effective Nonparametric Density Estimators in Several DimensionsThe Annals of Statistics, 1985
- Smoothing noisy data with spline functionsNumerische Mathematik, 1985
- A Fast and Efficient Cross-Validation Method for Smoothing Parameter Choice in Spline RegressionJournal of the American Statistical Association, 1984
- Algorithms for Computing the Sample Variance: Analysis and RecommendationsThe American Statistician, 1983
- Optimal Smoothing of Noisy Data Using Spline FunctionsSIAM Journal on Scientific and Statistical Computing, 1981
- Sur le choix du param tre d'ajustement dans le lissage par fonctions splineNumerische Mathematik, 1980
- Algorithm AS 52: Calculation of Power Sums of Deviations about the MeanJournal of the Royal Statistical Society Series C: Applied Statistics, 1972