Bayesian inference and variable selection in normal and binomial regression models with applications in medical research

Postgraduate Thesis uoadl:1314686 75 Read counter

Unit:
Κατεύθυνση Βιοστατιστική
Library of the School of Health Sciences
Deposit date:
2014-02-11
Year:
2013
Author:
Πατέρας Κωνσταντίνος
Supervisors info:
Ι. Ντζούφρας, Δ. Καρλής, Λ. Μελιγκοτσίδου
Original Title:
Bayesian inference and variable selection in normal and binomial regression models with applications in medical research
Languages:
English
Translated title:
Μπεϋζιανή συμπερασματολογία και επιλογή μεταβλητών σε γραμμικα και διωνυμικά μοντέλα με εφαρμογές στην Ιατρική
Summary:
In this Master Thesis we will deal with full Bayesian inference in both normal
and bino-
mial regression models when there is doubt about the structure of the linear
combination
and the parameters of the model. There will be a review of the relative
methodology
and emphasis will be placed in possible alternatives a priori distributions
which can be
used in such type of problems. The usage of advanced MCMC algorithms for a
priori
estimation of parameters, variable selection. In the frame of this Master
thesis synopsis,
application and comparison of existing code and programs in R and WinBUGS
environ-
ments for variable selection and model averaging will be presented. The
methodology
will be applied in both simulated and medical data with emphasis on those
derived from
the European Health Interview Survey (EHIS) 2009 held in Greece.
Keywords:
Bayesian, Variable Selection, Linear, Binomial, MCMC
Index:
Yes
Number of index pages:
xvi, xvii, xviii, xix, xx, xxi, 93, 94, 95, 96, 97, 98
Contains images:
Yes
Number of references:
169
Number of pages:
XXI,135

document.pdf
1 MB
File access is restricted only to the intranet of UoA.

 


attachments.zip
22 bytes
File access is restricted.