Nonparametric, isotonic discriminant procedures
成果类型：
Article
署名作者：
Dykstra, R; Hewett, J; Robertson, T
署名单位：
University of Iowa; University of Missouri System; University of Missouri Columbia
刊物名称：
BIOMETRIKA
ISSN/ISSBN：
0006-3444
DOI：
10.1093/biomet/86.2.429
发表日期：
1999
页码：
429438
关键词：
摘要：
In many situations it is desirable that prediction/discriminant rules should satisfy ordering properties with respect to the explanatory variables. Here, optimal rules under both L-1 and L-2 type loss functions are characterised under the restriction that the rules be isotonic with respect to the partial ordering on the explanatory variables. An example is given where college grade-point average is predicted based upon high school rank and American College Testing scores.
来源URL：