Do not adjust coefficients in Shapley value regression

被引：6

作者：

Groemping, Ulrike ^{[1
]}

Landau, Sabine ^{[2
]}

机构：

[1] Beuth Hsch Tech Berlin, Berlin, Germany

[2] Kings Coll London, Inst Psychiat, London, England

来源：

APPLIED STOCHASTIC MODELS IN BUSINESS AND INDUSTRY | 2010年 / 26卷 / 02期

关键词：

multiple regression; Shapley value regression; multicollinearity; Lord's paradox; suppression; RELATIVE IMPORTANCE; SMOKER;

D O I：

10.1002/asmb.773

中图分类号：

C93 [管理学]; O22 [运筹学];

学科分类号：

070105 ; 12 ; 1201 ; 1202 ; 120202 ;

摘要：

Shapley value regression consists of assessing relative importance and accordingly adjusting regression coefficients. It is argued that adjustment of coefficients is unnecessary and even misleading for practically relevant situations. Examples are given, and an alternative procedure is proposed for situations for which the coefficients are requested to have a certain sign. Copyright (C) 2009 John Wiley & Sons, Ltd.

引用

页码：194 / 202

页数：9

共 34 条

[1] Arah OA, 2008, EMERG THEMES EPIDEMI, V5, DOI 10.1186/1742-7622-5-5
[2] BECKLAKE MR, 1990, RESPIRATION, V57, P137
[3] DOMINANCE ANALYSIS - A NEW APPROACH TO THE PROBLEM OF RELATIVE IMPORTANCE OF PREDICTORS IN MULTIPLE-REGRESSION
BUDESCU, DV
[J]. PSYCHOLOGICAL BULLETIN, 1993, 114 (03) : 542 - 551
[4] Customer satisfaction analysis: Identification of key drivers
Conklin, M
Powaga, K
Lipovetsky, S
[J]. EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2004, 154 (03) : 819 - 827
[5] MULTIPLE REGRESSION IN PSYCHOLOGICAL RESEARCH AND PRACTICE
DARLINGTON, RB
[J]. PSYCHOLOGICAL BULLETIN, 1968, 69 (03) : 161 - +
[6] Imposing inequality restrictions: efficiency gains from economic theory
Dorfman, JH
McIntosh, CS
[J]. ECONOMICS LETTERS, 2001, 71 (02) : 205 - 209
[7] Feldman B., 2005, RELATIVE IMPOR UNPUB
[8] Gourieroux C., 1995, Statistics and Econometric Models, V2
[9] GREENBERG DH, 1970, R579OEO RAND CORP
[10] GRENVILLE A, 2007, EMOTIONAL REACTIONS

← 1 2 3 4 →