Multiple comparison procedures applied to model selection

被引：62

作者：

Pizarro, J

Guerrero, E

Galindo, PL ^{[1
]}

机构：

[1] Univ Cadiz, Dept Lenguajes & Sistemas Informat, Cadiz, Spain

[2] Univ Cadiz, Inteligencia Artificial Grupo Sistemas Inteligent, Cadiz, Spain

来源：

NEUROCOMPUTING | 2002年 / 48卷

关键词：

model selection; multiple comparison procedures; generalization; network size; problem complexity;

D O I：

10.1016/S0925-2312(01)00653-1

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This paper presents a new approach to model selection based on hypothesis testing. We first describe a procedure to generate different scores for any candidate model from a single sample of training data and then discuss how to apply multiple comparison procedures (MCP) to model selection. MCP statistical tests allow us to compare three or more groups of data while controlling the probability of making at least one Type I error. The complete procedure is illustrated on several model selection tasks, including the determination of the number of hidden units for feed-forward neural networks and the number of kernels for RBF networks. (C) 2002 Elsevier Science B.V. All rights reserved.

引用

页码：155 / 173

页数：19

共 35 条

[1] NEW LOOK AT STATISTICAL-MODEL IDENTIFICATION [J].