NONPARAMETRIC-ESTIMATION VIA EMPIRICAL RISK MINIMIZATION

被引：112
|
作者：
LUGOSI, G [1 ]
ZEGER, K [1 ]
机构：
[1] UNIV ILLINOIS,DEPT ELECT & COMP ENGN,COORDINATED SCI LAB,URBANA,IL 61801
来源：
IEEE TRANSACTIONS ON INFORMATION THEORY | 1995年 / 41卷 / 03期
基金：
美国国家科学基金会;
关键词：
REGRESSION ESTIMATION; NONPARAMETRIC ESTIMATION; CONSISTENCY; PATTERN RECOGNITION; NEURAL NETWORKS; SERIES METHODS; SIEVES;
D O I：
10.1109/18.382014
中图分类号：
TP [自动化技术、计算机技术];
学科分类号：
0812 ;
摘要：
A general notion of universal consistency of nonparametric estimators is introduced that applies to regression estimation, conditional median estimation, curve fitting, pattern recognition, and learning concepts. General methods for proving consistency of estimators based on minimizing the empirical error are shown, In particular, distribution-free almost sure consistency of neural network estimates and generalized linear estimators is established.
引用
收藏
页码：677 / 687
页数：11
相关论文
共 50 条
12345