Interpolation methods for adapting to sparse design in nonparametric regression

Research Report SRR95-021

Interpolation methods for adapting to sparse design in nonparametric regression

Peter Hall, Berwin A. Turlach

Abstract: We suggest interpolation methods for overcoming the problem of sparse design in local linear smoothing. They are based on simple rules, determined by the kernel and bandwidth, for deciding when and were pseudo design points should be added to augment the original design sequence. New ordinates for the added design points are computed by simple interpolation, and then local linear smoothing is applied directly to the expanded data set. The method is competitive with alternatives, for example those involving ridge regression, on grounds of both simplicity and performance.

This service is maintained by the Mathematical Sciences Institute (MSI)
Comments to webmaster@maths.anu.edu.au URL: http://wwwmaths.anu.edu.au/