IAP-Workshop 2009

Modeling Association and Dependence in Complex Data

Leuven, November 19-20, 2009

IAP BELSPO

Location:

Theromotechnisch Instituut, Kasteelpark Arenberg 41, 3001 Heverlee
In order to find the location easily, a plan is available for download.

Workshop topics:

  • WP1: Multivariate data with qualitative constraints
  • WP2: Temporally and spatially related data
  • WP3: Incomplete data
  • WP4: Data with latent heterogeneity
  • WP5: High-dimensional and compound data
  • WP6: Other

Keynote lectures:

  • Gerda Claeskens (Katholieke Universiteit Leuven, Belgium): Model testing and selection.

    Abstract : We will first discuss some specific model selection and model averaging methods. In particular we will explain the concept of focussed model selection. Sometimes a model is good for estimating a certain parameter, but not for another one. The focussed information criterion is tailored to select the best model for estimating your focus.
    We explain how to construct and use it in various situations. Model averaging amounts to estimation via a number of possible models and forming a weighted average of the resulting estimators. One example of model averaged estimators are estimators in a selected model. Model averaging clearly shows that ignoring model selection in any further inference in the selected models can lead to dramatic wrong conclusions. In the second part of the presentation, goodness-of-fit tests in linear mixed models will be discussed. Linear mixed models with both fixed and random effects are most often estimated on the assumption that the random effects have a normal distribution. The goodness-of-fit tests that we study are used to formally test the hypothesis that the random effects and/or the errors are normally distributed. The proposed tests are nonparametric and based on the order selection idea. They are designed to detect virtually any alternative to normality. In case of rejection of the null hypothesis, the nonparametric estimation method that is used in the construction of the test provides an estimator of the alternative distribution.


  • Timothy Hanson (University of Minnesota, U.S.A.): Bayesian survival analysis: An overview of models and methods.
    Abstract : In this talk I will provide an overview of approaches for analyzing time-to-event data using semiparametric and nonparametric models. Popular models including proportional hazards, accelerated failure time, proportional odds, additive hazards, and proportional mean residual life will be discussed, along with common parametric and nonparametric priors on the baseline distribution. Nonparametric priors include the gamma process, beta process, Dirichlet process mixtures, Polya trees, penalized splines and extensions of these. Then I will discuss various model generalizations including time dependent covariates, joint longitudinal and survival modeling, various frailty structures, cure models, and completely nonparametric dependent process approaches. Various models will be fit and illustrated in several software packages and languages including WinBUGS, BayesX, DPpackage, and FORTRAN.

Contributed presentations: 

Submission of abstract is closed.

Pre-conference short course (17-18 November):

`Non-parametric Bayesian statistics’
by Timothy Hanson (University of Minnesota, U.S.A.)  and Alejandro Jara Vallejos (Universidad de Concepción, Chile).

A short description of the course can be downloaded here. A course schedule can be found here. The short course will take place in the Thermotechnisch Instituut, Kasteelpark Arenberg 41, 3001 Heverlee, room 01.04 (leslokaal machinezaal). In order to find the location easily, a plan is available for download.

Program:

A summary of the workshop program can be found below. Full details with abstracts can be found in the program book.

Thursday, November 19

Friday, November 20

8.30 – 9.00

Registration

 

 

9.00 – 9.15

Welcome & Network introduction

9.00 – 10.15

Keynote session:

 

I. Van Keilegom

 

G. Claeskens (part II)

9.15 – 10.15

Keynote session:

 

 

G. Claeskens (part I)

 

10.15 – 11.25

WP2 contributions (4 papers)

10.15 – 11.25

WP3 contributions (4 papers)

 

Chair: I. Van Keilegom

 

Chair: G. Molenberghs

11.25 – 11.50

Coffee

11.25 – 11.50

Coffee

11.50 – 13.15

WP1 contributions (5 papers)

11.50 – 13.15

WP4 contributions (5 papers)

 

Chair: I. Gijbels

 

Chair: G. Verbeke

13.15 – 14.00

Lunch

13.15 – 14.00

Lunch

14.00 – 15.40

WP5 contributions (6 papers)

14.00 – 15.40

WP6 contributions (6 papers)

 

Chair: L. Duchateau

 

Chair: E. Lesaffre

15.40 – 16.00

Coffee

15.40 – 16.00

Coffee

16.00 – 17.15

Keynote session:

16.00 – 17.00

Keynote session:

 

T. Hanson (part I)

 

T. Hanson (part II)

17.30 – 19.00

Administrative meeting with BELSPO (closed meeting)

17.00 – . . .

Closing reception

 

List of paritcipants:

A list of participants can be found here (XLS)