Bayesian model comparison: Log scores and DIC

Milovan Krnjajić; David Draper

doi:10.1016/j.spl.2014.01.023

Bayesian model comparison: Log scores and DIC

Milovan Krnjajić, David Draper

Mathematics

Research output: Contribution to a Journal (Peer & Non Peer) › Article › peer-review

10 Citations (Scopus)

Abstract

We investigate Bayesian approaches to answering the frequently-arising model-specification question (Q:) Is model _{M j} better than Mj'? We contrast two Bayesian model-comparison methods-log scores and D I C-on their ability to correctly discriminate between fixed-effects Poisson (FEP) and random-effects Poisson (REP) sampling models.

Original language	English
Pages (from-to)	9-14
Number of pages	6
Journal	Statistics and Probability Letters
Volume	88
Issue number	1
DOIs	https://doi.org/10.1016/j.spl.2014.01.023
Publication status	Published - May 2014

Keywords

Bayesian model discrimination
Calibration
Cross-validation log score
D I C
Full-sample log score
Prediction

Access to Document

10.1016/j.spl.2014.01.023

Cite this

@article{1150e1b1985f46669ee49cd38f1533d4,

title = "Bayesian model comparison: Log scores and DIC",

abstract = "We investigate Bayesian approaches to answering the frequently-arising model-specification question (Q:) Is model M j better than Mj'? We contrast two Bayesian model-comparison methods-log scores and D I C-on their ability to correctly discriminate between fixed-effects Poisson (FEP) and random-effects Poisson (REP) sampling models.",

keywords = "Bayesian model discrimination, Calibration, Cross-validation log score, D I C, Full-sample log score, Prediction",

author = "Milovan Krnjaji{\'c} and David Draper",

year = "2014",

month = may,

doi = "10.1016/j.spl.2014.01.023",

language = "English",

volume = "88",

pages = "9--14",

journal = "Statistics and Probability Letters",

issn = "0167-7152",

publisher = "Elsevier B.V.",

number = "1",

}

TY - JOUR

T1 - Bayesian model comparison

T2 - Log scores and DIC

AU - Krnjajić, Milovan

AU - Draper, David

PY - 2014/5

Y1 - 2014/5

N2 - We investigate Bayesian approaches to answering the frequently-arising model-specification question (Q:) Is model M j better than Mj'? We contrast two Bayesian model-comparison methods-log scores and D I C-on their ability to correctly discriminate between fixed-effects Poisson (FEP) and random-effects Poisson (REP) sampling models.

AB - We investigate Bayesian approaches to answering the frequently-arising model-specification question (Q:) Is model M j better than Mj'? We contrast two Bayesian model-comparison methods-log scores and D I C-on their ability to correctly discriminate between fixed-effects Poisson (FEP) and random-effects Poisson (REP) sampling models.

KW - Bayesian model discrimination

KW - Calibration

KW - Cross-validation log score

KW - D I C

KW - Full-sample log score

KW - Prediction

UR - https://www.scopus.com/pages/publications/84894099761

U2 - 10.1016/j.spl.2014.01.023

DO - 10.1016/j.spl.2014.01.023

M3 - Article

AN - SCOPUS:84894099761

SN - 0167-7152

VL - 88

SP - 9

EP - 14

JO - Statistics and Probability Letters

JF - Statistics and Probability Letters

IS - 1

ER -

Bayesian model comparison: Log scores and DIC

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this