Model selection using information criteria, but is the “best” model any good?

Journal of Applied Ecology - Tập 55 Số 3 - Trang 1441-1444 - 2018
Ralph Mac Nally1,2, Richard P. Duncan2, James R. Thomson3, Jian D. L. Yen4
1Department of Ecology, Environment and Evolution, La Trobe University, Bundoora, Australia
2Institute for Applied Ecology The University of Canberra Bruce ACT Australia
3Department of Environment, Land, Water and Planning Arthur Rylah Institute for Environmental Research Melbourne Vic. Australia
4School of BioSciences, The University of Melbourne, Parkville, VIC, Australia

Tóm tắt

Abstract Information criteria (ICs) are used widely for data summary and model building in ecology, especially in applied ecology and wildlife management. Although ICs are useful for distinguishing among rival candidate models, ICs do not necessarily indicate whether the “best” model (or a model‐averaged version) is a good representation of the data or whether the model has useful “explanatory” or “predictive” ability. As editors and reviewers, we have seen many submissions that did not evaluate whether the nominal “best” model(s) found using IC is a useful model in the above sense. We scrutinized six leading ecological journals for papers that used IC to compare models. More than half of papers using IC for model comparison did not evaluate the adequacy of the best model(s) in either “explaining” or “predicting” the data. Synthesis and applications. Authors need to evaluate the adequacy of the model identified as the “best” model by using information criteria methods to provide convincing evidence to readers and users that inferences from the best models are useful and reliable.

Từ khóa


Tài liệu tham khảo

10.1080/02664769823250

10.1038/nature.2016.19503

10.1177/0049124104268644

10.1002/eap.1438

10.1016/j.biocon.2017.03.006

10.1016/j.ecolind.2017.09.020

10.1002/smj.582

10.1002/jwmg.21253

Long J. S., 1997, Regression models for categorical and limited dependent variables

10.1002/jwmg.21175

10.1007/978-1-4899-3242-6

10.1002/eap.1479

10.1111/j.2041-210x.2012.00261.x

R Core Team, 2016, R: A language and environment for statistical computing

10.1111/ecog.02881

10.1111/j.1365-2664.2005.01002.x

Steyerberg E., 2008, Clinical prediction models: A practical approach to development, validation, and updating

10.1007/s00265-010-1037-6

10.1007/s11222-016-9696-4

Watanabe S., 2010, Asymptotic equivalence of Bayes cross validation and widely applicable information criterion in singular learning theory, Journal of Machine Learning Research, 11, 3571

Watanabe S., 2013, A widely applicable Bayesian information criterion, Journal of Machine Learning Research, 14, 867

10.1111/2041-210X.12290

10.1111/2041-210X.12577

Thông tin
Thông tin xuất bản

Journal of Applied Ecology

Tập 55 Số 3

1441-1444

Thông tin tác giả