Please use this identifier to cite or link to this item:
|Title:||Variational approximation for heteroscedastic linear models and matching pursuit algorithms|
|Authors:||Nott, D.J. |
|Keywords:||Bayesian model selection|
|Source:||Nott, D.J., Tran, M.-N., Leng, C. (2012-03). Variational approximation for heteroscedastic linear models and matching pursuit algorithms. Statistics and Computing 22 (2) : 497-512. ScholarBank@NUS Repository. https://doi.org/10.1007/s11222-011-9243-2|
|Abstract:||Modern statistical applications involving large data sets have focused attention on statistical methodologies which are both efficient computationally and able to deal with the screening of large numbers of different candidate models. Here we consider computationally efficient variational Bayes approaches to inference in high-dimensional heteroscedastic linear regression, where both the mean and variance are described in terms of linear functions of the predictors and where the number of predictors can be larger than the sample size. We derive a closed form variational lower bound on the log marginal likelihood useful for model selection, and propose a novel fast greedy search algorithm on the model space which makes use of one-step optimization updates to the variational lower bound in the current model for screening large numbers of candidate predictor variables for inclusion/exclusion in a computationally thrifty way. We show that the model search strategy we suggest is related to widely used orthogonal matching pursuit algorithms for model search but yields a framework for potentially extending these algorithms to more complex models. The methodology is applied in simulations and in two real examples involving prediction for food constituents using NIR technology and prediction of disease progression in diabetes. © 2011 Springer Science+Business Media, LLC.|
|Source Title:||Statistics and Computing|
|Appears in Collections:||Staff Publications|
Show full item record
Files in This Item:
There are no files associated with this item.
checked on Mar 7, 2018
WEB OF SCIENCETM
checked on Feb 5, 2018
checked on Mar 11, 2018
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.