University of Liege | Version française
Study programmes 2007-2008Last update : 7/05/2008
STAT0723-1  Modèles linéaires
Duration :  30h Th, 30h Pr
Credits/ECTS :  
Master in Statistics : General, Research focus, 1st year6
Master in Statistics : General, Professional focus, 1st year6
Holder(s) :  Paul Gérard
Language :  Langue française
Course contents :  The course is mainly devoted to a detailed study of the normal general linear model. The main particular models are examined: ANOVA I and ANOVA II (fixed, random and mixed models), factorial designs, multiple regression.

Maximum likelihood estimation and introduction to the generalized linear model.

Logistic regression and log-linear analysis are presented.
Course objective :  To learn the fundamentals of probability theory and statistical data analysis. To get accustomed with the notion of statistical model and to put a model into practise by using a statistical software.
Prerequisites :  Knowledge of probability theory.

Essentials about statistical analysis.
Workshops :  Treatment of experimental data
Written notes :  A summary of the course is available.

Reference Books :

Applied linear statistical models. J.Neter, M.H.Kutner, Ch.J.Nachtsheim, W.Wasserman. Wcb McGraw-Hill

Applied multivariate data analysis volume1: regression and experimental design. J.D.Jobson

Applied logistic regression. D.W.Hosmer, S.Lemeshow. Wiley series in probability and statistics.

Multivariate statistical modelling based on generalized linear models. L.Fahrmeir, G.Tutz. Springer series in statistics
Assessment :  Oral examination

Written examination

Statistical analysis report on a set of experimental data
Contacts :  Paul GERARD

Institut de mathématique

Grande traverse, 12

Sart-Tilman , B-4000 Liège

Tél.: 00-32-(0)366.93.84

Mail: paul.gerard@ulg.ac.be


imageHome
imageSearch by Faculty
imageSearch by teacher
imageSearch by course code and title

Students and Studies Administration - Academic Affairs - Contact : Monique Marcourt, General Director for Education and Training - Developed by SEGI