Alternative regression methods are not considered in Murtaugh (2009) or by ecologists in general

被引：44

作者：

Dahlgren, Johan P. ^{[1
]}

机构：

[1] Stockholm Univ, Dept Bot, SE-10691 Stockholm, Sweden

来源：

ECOLOGY LETTERS | 2010年 / 13卷 / 05期

关键词：

AIC; all subsets; BIC; lasso; ridge regression; shrinkage; stepwise multiple regression; variable selection; SELECTION; INFERENCE; VARIABLES; NOISE;

D O I：

10.1111/j.1461-0248.2010.01460.x

中图分类号：

Q14 [生态学（生物生态学）];

学科分类号：

071012 ; 0713 ;

摘要：

P>Murtaugh (2009) recently illustrated that all subsets variable selection is very similar to stepwise regression. This, however, does not necessarily mean both methods are useful. On the contrary, the same problems with overfitting should apply. Ecologists should, if model building is indeed necessary, consider more reliable regression methods now available.

引用

页码：E7 / E9

页数：3

共 16 条

[1]

COPAS JB, 1983, J R STAT SOC B, V45, P311

[2]

Draper N., 1981, Applied Regression Analysis, V7th

[3] FREQUENCY OF SELECTING NOISE VARIABLES IN SUBSET REGRESSION-ANALYSIS - A SIMULATION STUDY [J].

FLACK, VF ;

CHANG, PC .

AMERICAN STATISTICIAN, 1987, 41 (01) :84-86

[4] L1 Penalized Estimation in the Cox Proportional Hazards Model [J].

Goeman, Jelle J. .

BIOMETRICAL JOURNAL, 2010, 52 (01) :70-84

[5]

Harrell FE., 2001, Regression Modeling Strategies: with Applications to Linear Models, Logistic Regression, and Survival Analysis, V608, DOI DOI 10.2147/

[6] RIDGE REGRESSION - BIASED ESTIMATION FOR NONORTHOGONAL PROBLEMS [J].

HOERL, AE ;

KENNARD, RW .

TECHNOMETRICS, 1970, 12 (01) :55-&

[7] THE IMPACT OF MODEL SELECTION ON INFERENCE IN LINEAR-REGRESSION [J].

HURVICH, CM ;

TSAI, CL .

AMERICAN STATISTICIAN, 1990, 44 (03) :214-217

[8] Stepwise Model Fitting and Statistical Inference: Turning Noise into Signal Pollution [J].

Mundry, Roger ;

Nunn, Charles L. .

AMERICAN NATURALIST, 2009, 173 (01) :119-123

[9] Performance of several variable-selection methods applied to real ecological data [J].

Murtaugh, Paul A. .

ECOLOGY LETTERS, 2009, 12 (10) :1061-1068

[10] A Review of Bayesian Variable Selection Methods: What, How and Which [J].

O'Hara, R. B. ;

Sillanpaa, M. J. .

BAYESIAN ANALYSIS, 2009, 4 (01) :85-117

← 1 2 →