MODEL AVERAGING BASED ON KULLBACK-LEIBLER DISTANCE. Academic Article

abstract

• This paper proposes a model averaging method based on Kullback-Leibler distance under a homoscedastic normal error term. The resulting model average estimator is proved to be asymptotically optimal. When combining least squares estimators, the model average estimator is shown to have the same large sample properties as the Mallows model average (MMA) estimator developed by Hansen (2007). We show via simulations that, in terms of mean squared prediction error and mean squared parameter estimation error, the proposed model average estimator is more efficient than the MMA estimator and the estimator based on model selection using the corrected Akaike information criterion in small sample situations. A modified version of the new model average estimator is further suggested for the case of heteroscedastic random errors. The method is applied to a data set from the Hong Kong real estate market.

authors

• Carroll, Raymond

published proceedings

• Stat Sin

author list (cited authors)

• Zhang, X., Zou, G., & Carroll, R. J.

citation count

• 47

complete list of authors

• Zhang, Xinyu||Zou, Guohua||Carroll, Raymond J

publication date

• October 2015

publisher

• Statistica Sinica (Institute of Statistical Science)  Publisher

published in

• STATISTICA SINICA  Journal

keywords

• Akaike Information
• Kullback-leibler Distance
• Model Averaging
• Model Selection
• Prediction

Digital Object Identifier (DOI)

• 10.5705/ss.2013.326

start page

• 1583

end page

• 1598

volume

• 25

issue

• 4

URL

• http%3A%2F%2Fdx.doi.org%2F10.5705%2Fss.2013.326