Model selection bias and Freedman’s paradox

By P. Lukacs, K. Burnham, David R. Anderson. 2010

331 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/d26f000897a6d36dd51a1e50ecde5c7a8c59ede6

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

Citations in the corpus

Multimodel inference in ecology and evolution: challenges and solutions (2011, neutral): “these predictors (shrinkage towards zero) (Lukacs et al., 2010).”
Towards a Standard for Identifying and Managing Bias in Artificial Intelligence (2022, background): “Model selection bias also occurs when an explanatory variable has a weak relationship with the response variable [281].”
MMI: Multimodel inference or models with management implications? (2015, methodology): “When model averaging regression coefficients, one should include a zero coefficient for explanatory variables that are absent from a given model (Lukacs et al. 2010).”
A farewell to the sum of Akaike weights: The benefits of alternative metrics for variable importance estimations in model selection (2017, methodology): “Under the full model-averaging technique, pa - rameter estimates are multiplied by the Akaike weight of their corre - sponding model (Lukacs et al., 2010).”
Sperm number and velocity affect sperm competition success in the guppy (Poecilia reticulata) (2011, methodology): “To test the effect of the predictors, we used an information theoretic (IT) approach (Burnham and Anderson 2002; Lukacs et al. 2010).”
Evaluating a coupled discrete wavelet transform and support vector regression for daily and monthly streamflow forecasting (2014, background): “(10) can be calculated using the following full-model averaging form (Lukacs et al., 2009):”
of construction probability on word durations during spontaneous incremental sentence production (2012, methodology): “The motivation and mathematical foundations of this approach, as well as the criticism of the methods of single best model selection are exhaustively presented in Burnham and Anderson (2002, 2004) and Lukacs et al. (2010).”
Using information theory as a substitute for stepwise regression in ecology and behavior (2010, neutral): “2005), and even reference methodological papers have used all subset IT approaches (see, e.g., Whittingham et al. 2006; Lukacs et al. 2010).”
A brief guide to model selection, multimodel inference and model averaging in behavioural ecology using Akaike’s information criterion (2010, methodology): “The second approach to model averaging, known as fullmodel averaging (see Lukacs et al. 2009), should be employed in cases where high model selection uncertainty exists (i.e. the best AIC model is not strongly weighted); thus, inference needs to be based on all models in the candidate set.”