Entropic kernels for data smoothing

被引:1
|
作者
Bowden, Roger [1 ]
机构
[1] Kiwicap Res Ltd, Wellington 6012, New Zealand
关键词
Bandwidth; Data smoothing; Kernel regression; Locational entropy; Subjective probabilities;
D O I
10.1016/j.spl.2012.12.006
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
Data smoothing or regression kernels based on locational entropy embody the principle that observations towards the extremes of the chosen data window should provide less information than those at the midpoint. Weight patterns can be flexible, depending on the choice of prior information density. (C) 2012 Elsevier B.V. All rights reserved.
引用
收藏
页码:916 / 922
页数:7
相关论文
共 50 条