Forecasting seasonal time series data: a Bayesian model averaging approach

Vosseler, Alexander; Weber, Enzo

doi:10.1007/s00180-018-0801-3

Cited by 14 publications

(10 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Of course, BMA is not limited to these scenarios and can be applied whenever there is model uncertainty. Other examples of BMA applications include the estimation of effect size (Haldane, 1932), linear regression (Clyde, Ghosh, & Littman, 2011), assessment of the replicability of effects (Iverson, Wagenmakers, & Lee, 2010), prediction in time-series analysis (Vosseler & Weber, 2018), analysis of the causal structure in a brain network (Penny et al, 2010), structural equation modeling (Kaplan & Lee, 2016), factor analysis (Dunson, 2006), and correcting for publication bias using the precision-effect test and precision-effect estimate with standard errors (Carter & McCullough, 2018). In general, BMA reduces overconfidence, results in optimal predictions (under mild conditions), avoids threshold-based all-or-nothing decision making, and is relatively robust against model misspecification.…”

Section: Concluding Commentsmentioning

confidence: 99%

A Conceptual Introduction to Bayesian Model Averaging

Hinne

Gronau

Bergh

et al. 2020

Advances in Methods and Practices in Psychological Science

234

188

View full text Add to dashboard Cite

Many statistical scenarios initially involve several candidate models that describe the data-generating process. Analysis often proceeds by first selecting the best model according to some criterion and then learning about the parameters of this selected model. Crucially, however, in this approach the parameter estimates are conditioned on the selected model, and any uncertainty about the model-selection process is ignored. An alternative is to learn the parameters for all candidate models and then combine the estimates according to the posterior probabilities of the associated models. This approach is known as Bayesian model averaging (BMA). BMA has several important advantages over all-or-none selection methods, but has been used only sparingly in the social sciences. In this conceptual introduction, we explain the principles of BMA, describe its advantages over all-or-none model selection, and showcase its utility in three examples: analysis of covariance, meta-analysis, and network analysis.

show abstract

Section: Concluding Commentsmentioning

confidence: 99%

A Conceptual Introduction to Bayesian Model Averaging

Hinne

Gronau

Bergh

et al. 2020

Advances in Methods and Practices in Psychological Science

234

188

View full text Add to dashboard Cite

show abstract

“…Of course, BMA is not limited to these scenarios and can be applied whenever there is model uncertainty. Other examples of other BMA applications include the estimation of effect size (Haldane, 1932), linear regression (Clyde et al, 2011), replicability of effects (Iverson, Wagenmakers, & Lee, 2010), prediction in time series analysis (Vosseler & Weber, 2018), analysis of the causal structure between brain regions (Penny et al, 2010), structural equation modelling (Kaplan & Lee, 2016), factor analysis (Dunson et al, 2006) and the PET-PEESE decision in correcting for publication bias (Carter & McCullough, 2018). In general, BMA reduces overconfidence, results in optimal predictions (under mild conditions), avoids threshold-based all-or-nothing decision making and is relatively robust against model-misspecification.…”

Section: Concluding Commentsmentioning

confidence: 99%

A conceptual introduction to Bayesian Model Averaging

Hinne¹,

Gronau²,

Bergh³

et al. 2019

Preprint

View full text Add to dashboard Cite

Many statistical scenarios initially involve several candidate models that describe the data-generating process. Analysis often proceeds by first selecting the best model according to some criterion, and then learning about the parameters of this selected model. Crucially however, in this approach the parameter estimates are conditioned on the selected model, and any uncertainty about the model selection process is ignored. An alternative is to learn the parameters for all candidate models, and then combine the estimates according to the posterior probabilities of the associated models. The result is known as Bayesian model averaging (BMA). BMA has several important advantages over all-or-none selection methods, but has been used only sparingly in the social sciences. In this conceptual introduction we explain the principles of BMA, describe its advantages over all-or-none model selection, and showcase its utility for three examples: ANCOVA, meta-analysis, and network analysis.

show abstract

“…Of course, BMA is not limited to these scenarios and can be applied whenever there is model uncertainty. Other examples of BMA applications include the estimation of effect size (Haldane, 1932), linear regression (Clyde et al, 2011), replicability of effects (Iverson, Wagenmakers, & Lee, 2010), prediction in time series analysis (Vosseler & Weber, 2018), analysis of the causal structure between brain regions (Penny et al, 2010), structural equation modelling (Kaplan & Lee, 2016), factor analysis (Dunson et al, 2006) and the PET-PEESE decision in correcting for publication bias (Carter & McCullough, 2018). In general, BMA reduces overconfidence, results in optimal predictions (under mild conditions), avoids threshold-based allor-nothing decision making and is relatively robust against model-misspecification.…”

Section: Concluding Commentsmentioning

confidence: 99%

A conceptual introduction to Bayesian model averaging

Hinne¹,

Gronau²,

Bergh³

et al. 2019

Preprint

View full text Add to dashboard Cite

Many statistical scenarios initially involve several candidate models that describe the data-generating process. Analysis often proceeds by first selecting the best model according to some criterion, and then learning about the parameters of this selected model. Crucially however, in this approach the parameter estimates are conditioned on the selected model, and any uncertainty about the model selection process is ignored. An alternative is to learn the parameters for allcandidate models, and then combine the estimates according to the posterior probabilities of the associated models. The result is known as Bayesian model averaging (BMA). BMA has several important advantages over all-or-none selection methods, but has been used only sparingly in the social sciences. In this conceptual introduction we explain the principles of BMA, describe its advantages over all-or-none model selection, and showcase its utility for three examples: ANCOVA, meta-analysis, and network analysis.

show abstract

Forecasting seasonal time series data: a Bayesian model averaging approach

Cited by 14 publications

References 53 publications

A Conceptual Introduction to Bayesian Model Averaging

A Conceptual Introduction to Bayesian Model Averaging

A conceptual introduction to Bayesian Model Averaging

A conceptual introduction to Bayesian model averaging

Contact Info

Product

Resources

About