Article ID Journal Published Year Pages File Type
1151725 Statistics & Probability Letters 2014 7 Pages PDF
Abstract

For order qq kernel density estimators we show that the constant bqbq in bias=bqhq+o(hq)bias=bqhq+o(hq) can be made arbitrarily small, while keeping the variance bounded. A data-based selection of bqbq is presented and Monte Carlo simulations illustrate the advantages of the method.

Keywords
Related Topics
Physical Sciences and Engineering Mathematics Statistics and Probability
Authors
, , , ,