Please use this identifier to cite or link to this item: http://scholarbank.nus.edu.sg/handle/10635/135802
Title: IMPROVED BAYESIAN MODEL SPECIFICATION VIA ROBUST METHODS AND CHECKING FOR PRIOR-DATA CONFLICT
Authors: WANG XUEOU
Keywords: Variational Bayesian, Bayesian inference, Approximate Bayesian computation, History matching, Prior-data conflict, Model checking
Issue Date: 20-Dec-2016
Source: WANG XUEOU (2016-12-20). IMPROVED BAYESIAN MODEL SPECIFICATION VIA ROBUST METHODS AND CHECKING FOR PRIOR-DATA CONFLICT. ScholarBank@NUS Repository.
Abstract: In high-dimensional data analysis, variational Bayes has been a popular and adaptable computational method in making Bayesian inference. In this thesis, we implement variational Bayes method as an efficient computational tool to work on the following problems of interest. The first one is to make a robust regression estimation while at the same time identifying multiple outliers in "wide data", where we have more variables than the sample size. We show that with a Horseshoe+ prior distribution and a sparse signal regression model, we can effectively identify signals and outliers. The other problem is in hierarchical Bayesian analysis, specification of the prior distribution can be a technical issue, especially in complex models. A prior data conflict can arise if a prior distribution is not specified properly. We propose a novel method to make a reasonable prior distribution selection. Besides we also develop a new approach to check prior-data conflict in hierarchical Bayesian analysis.
URI: http://scholarbank.nus.edu.sg/handle/10635/135802
Appears in Collections:Ph.D Theses (Open)

Show full item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
WangXO.pdf9.17 MBAdobe PDF

OPEN

NoneView/Download

Page view(s)

13
checked on Jan 21, 2018

Download(s)

21
checked on Jan 21, 2018

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.