An introduction to ensemble methods for data analysis

被引:94
作者
Berk, Richard A. [1 ]
机构
[1] Univ Calif Los Angeles, Los Angeles, CA 90024 USA
关键词
CART; data analysis; algorithmic methods; ensemble methods;
D O I
10.1177/0049124105283119
中图分类号
O1 [数学]; C [社会科学总论];
学科分类号
03 ; 0303 ; 0701 ; 070101 ;
摘要
This article provides an introduction to ensemble statistical procedures as a special case of algorithmic methods. The discussion begins with classification and regression trees (CART) as a didactic device to introduce many of the key issues. Following the material on CART is a consideration of cross-validation, bagging, random forests, and boosting. Major points are illustrated with analyses of real data.
引用
收藏
页码:263 / 295
页数:33
相关论文
共 31 条
[1]   Developing a practical forecasting screener for domestic violence incidents [J].
Berk, RA ;
He, Y ;
Sorenson, SB .
EVALUATION REVIEW, 2005, 29 (04) :358-383
[2]  
BERK RA, IN PRESS QUANTITATIV
[3]  
BERK RA, 2003, UNPUB ENSEMBLE PROCE
[4]   Random forests [J].
Breiman, L .
MACHINE LEARNING, 2001, 45 (01) :5-32
[5]   Random forests [J].
Breiman, L .
MACHINE LEARNING, 2001, 45 (01) :5-32
[6]   Statistical modeling: The two cultures [J].
Breiman, L .
STATISTICAL SCIENCE, 2001, 16 (03) :199-215
[7]  
BREIMAN L, 1984, CLASSIFICASTION REGR
[8]  
BREIMAN L, 2001, WALD LECT 1 MACHINE
[9]  
Bühlmann P, 2002, ANN STAT, V30, P927
[10]  
Cristianini N., 2000, Intelligent Data Analysis: An Introduction, DOI 10.1017/CBO9780511801389