Statistics Home | People | Degrees | Papers | Research | Events | Other FCMS Subjects
 

UpComing Seminars

Recent Seminars

Assoc. Prof. Simon Wilson
Bayesian Source Separation of Cosmic Microwave Background

Prof Merrilyn Goos; Assoc Prof Sergiy Klymchuk; Prof Mike Thomas
Transition from School to University Education in Mathematics: New Zealand and Australia Perspectives

Assoc. Prof. Ross Ihaka
"R Functions: Things Your Mother Didn't Tell You About" followed by the inaugural meeting of the Waikato R Users Group.

Ben Stevenson
Predicting daily fishing success: The assessment of lunar and indigenous fishing calendars

Dr Zazli Chik
The influence of outliers on comparing stationary stochastic signals using the randomisation test method.

Dr Vladimir Obolonkin
Belarus and Chernobyl; Facts and Thoughts from a Statistician-Insider

Dr Nicholas J. Horton

Much ado about nothing: a review of the state of the art of incomplete data

Professor Rosemary Bailey
EXPERIMENTS IN RECTANGULAR AREAS: RESTRICTED RANDOMIZATION OF ROW-COLUMN DESIGNS

Professor Ray Chambers

Maximum Likelihood under Informative Sampling

Dr David Scott
The Hyperbolic and Related Distributions

Dr Russell Millar
ASSESSMENT OF HIERARCHICAL MODELS FOR COUNT DATA

Dr Rolf Turner
Direct Maximization of the Likelihood of a Hidden Markov Model

Dr Murray Jorgensen

Where have all the Young Men gone?

Professor Bruce Weir

THE BIRTHDAY PROBLEM and DNA PROFILES

Professor William J Reed

Normal-Laplace Distributions and their Applications

Hadley Wickham
Interactive Graphics

John Leathwick (NIWA) and Jane Elith (University of Melbourne)
Boosted Regression Trees - a major advance in statistical modelling tools

Dr Yong Wang
Extending the Fisher scoring and Gauss-Newton methods for complex statistical optimisation problems

Dr Thomas Yee
A Framework for Regression

Dr Murray Jorgensen
The Development of Statistics in New Zealand

Rohan Maheswaran
A New Approach to Robust Estimation

Dr Murthy Mittinty
Nearest Neighbour Imputation

Richard Penny
Statistics New Zealand and the Official Statistics System - Data Collection and Access

Dr James Curran
A MCMC method for resolving two person DNA mixtures

Dr Lyn Hunt
Using multiple choice questions as an assessment device for statistical thinking

Associate Professor David Whitaker
The state of the art in the generation of efficient statistical designs

Professor J A John
Design, GenStat and CycDesigN

Peter Mullins
Using Excel for inspecting segments of a time series of reasonable length

Dr Beatrix Jones
Fitting and interpreting sparse Gaussian graphical models for high dimensional data

Professor Chris Triggs
Establishing Identity

Dr Michael Stuart
Mathematical thinking versus statistical thinking; redressing the balance in statistical teaching

Dr Claire Jordan
CART, PPM's and PBRF

Professor George Styan
Issai Schur (1875-1941) and the early development of the Schur complement, SRP-1Q: photographs, documents and biographical remarks.

Dr Daniel C. I. Walsh
The Importance Sampling Hough Transform

Jan Bulla
Introduction to Hidden Markov Models

Angelika van der Linde
General measures of variability and dependence for multivariate continuous distributions.

Professor Rosemary Bailey
Hasse diagrams to describe the structure of designed experiments

Professor Jacques Poot
On the use of meta-analysis in economics

Professor J A John
Inverse of the Information Matrix

Dr David Johnson
Teaching Statistics with Microsoft Excel

Professor Peter Davis
Social facts and social stats: applications in quantitative sociology

Dr James Curran
Some issues surrounding the interpretation of Low Copy Number (LCN) DNA evidence.

Dr Murray Jorgensen
Estimating a density of microbial densities

Harold V. Henderson
Dynamic Graphics in Statistical Consulting

Professor David M Ryan
Robustness Issues in Aircrew Tours of Duty Optimisation

Dr Ray Hoare
Why do we need ANOTHER statistics program?

Dr Colin Aitken
Evaluation of Trace Evidence in the form of Multivariate Data

Dr Michael Bulmer
Virtual Worlds for Teaching Mathematics and Statistics

Graham McBride
Some Issues in Quantitative Health Risk Assessments for Waterborne Diseases

Professor Chris Wild
Regression Problems with Missing Data

Megan Jowsey (Project Coordinator) and Rachel Cunliffe (Web site/Publicity Developer)
CensusAtSchool NZ: a statistical experience for children

Nicholas T Longford
Stability of household income in European countries

Dr Neil Cox
What role does Excel have in the practice and teaching of statistics?

Dr Lyn Hunt
Unsupervised Learning from Incomplete Three-way Data using a Mixture Model Approach

Dr James Curran
Two Problems in Forensic Science

Ian Westbrooke
Meeting statistical needs in a conservation management organisation

Emlyn Williams
Statistical consulting and research in CSIRO Forestry and Forest Products

Carole Wright
Row-column Designs and their Contractions

Dave Saville
Multiple comparison procedures: consistency and family-size robustness

Associate Professor David Whitaker
Optimal Stratified Sampling Designs

Hans Hockey
Drug Development and Statistics

Robert G. Easterling*
Statistical Foundations for the Validation of Computer Models

Professor Keith J Worsley
Heritability random fields

Dr Niall Broekhuizen
Verification of an Individual-Based Model of the foraging of Southern Buller's Albatross by comparison with ship-board observations

Bryan Manly
The Assessment of Forestry on Northern Spotted Owls

Dr Robert G. Newcombe
Confidence Intervals for Proportions and Related Quantities

Dr Martin Upsdell
From Analysis of Means and Analysis of Coefficients to Analysis of Curves

Dr Mik Black
Statistical Issues in the Design and Analysis of Spotted Microarray Experiments

Dr Marti J. Anderson
Canonical analysis of principal coordinates: flexible constrained ordination

Dr Jennifer Brown
Designing an efficient adaptive cluster sample

Dr Judi McWhirter
Catching the Bayesian Wave: An MCMC Approach to Modelling Pulsatile Data

Jock MacKay
Some Issues in Variation Reduction

Dr M A Jorgensen
Using finite mixtures to robustify statistical models

Associate Professor Stefan Steiner
Seven Habits of Highly Effective Industrial Problem Solvers: An Overview of Statistical Engineering (Shanin Methods)

Professor J A John
Crossover Designs in Clinical Trials

Associate Professor David Whitaker
Integer Programming: Constraint formulation with binary variables

Dr Anthony Lehmann
GRASP: Generalized Regression Analysis and Spatial Prediction in Ecology

Professor Alastair Scott
Estimating Interviewer and Observer Effects for Binary Responses

Professor Les Foulds
Bookmobile Routing and Scheduling in Buskerud County, Norway

Associate Professor Andy Philpott
Optimisation in Electricity Pool Markets

Seminars
Recent Seminars
Dr Russell Millar - ASSESSMENT OF HIERARCHICAL MODELS FOR COUNT DATA

Dept Statistics, University of Auckland

2007-10-29

University of Waikato, Room I.1.09 (I block, 1st floor)

The deviance information criterion (DIC) was found to be a potentially dangerous tool for comparison between hierarchical models of count data and the results herein show that it has been used inappropriately in several recent publications that have employed these models. DIC was useful only when the likelihood was expressed at the subject level, and so it can be used for comparison between the Poisson and negative binomial.

In addition, despite zero-inflation being a form of mixture, the DIC also performed well for comparison of the Poisson and negative binomial with their zero-inflated counterparts.

However, DIC was not reliable for likelihoods expressed at the replicate level, and it can not be used to distinguish between Poisson-gamma (the negative binomial implemented at replicate level) and Poisson-lognormal models, or to assess whether these models require zero-inflation.

For example, when fitting Poisson-gamma and oisson-lognormal models to simulated Poisson-lognormal data, the oisson-gamma model always had lower DIC. Bayesian predictive checks (BPCs) were also investigated and were found to be extremely conservative. For example, under 100 simulations of the Poisson model fitted to Poisson data, the lower 5% quantile of the BPC p-value for goodness of fit was approximately 0.3.

  2007 FCMS. The University of Waikato - Te Whare Wananga o Waikato