Statistical Modelling

Faculteit Science and Engineering
Jaar 2017/18
Vakcode WISM-08
Vaknaam Statistical Modelling
Niveau(s) bachelor
Voertaal Engels
Periode semester II a

Uitgebreide vaknaam Statistical Modelling
Leerdoelen The student is able
1. to state and apply a taxonomy of Generalized Linear Models (GLM) in terms of measurement levels of the response, link function, design of (repeated) data collection related to the exponential family of distributions;
2. to mathematically derive and computationally implement maximum likelihood estimation of a GLM by iterative weighted least squares using the R programming environment;
3. to integrate the testing of medical/economic hypotheses to appropriate statistical model fitting as well as their selection by likelihood ratio or Akaike’s Information Criterion;
4. to evaluate the quality of model estimation in various applied scientific fields.
Omschrijving Taking the exponential family of distributions as a starting point, various generalized linear models are developed to model non-Gaussian response data. The aim is to describe this response in terms of a number of explanatory variables. Estimated models are evaluated and interpreted in order to test hypotheses from various scientific fields. In this course we will introduce estimation principles underlying generalized linear modeling. This course covers the following themes:
- Exponential family
- Maximum likelihood, Deviance, likelihood ratio
- Iterative weighted least squares
- Model Design
- Goodness of fit, residuals
- Linear, logistic (multinomial), Poisson regression, Loglinear models, Survival Analysis
Uren per week
Onderwijsvorm Hoorcollege (LC), Opdracht (ASM), Werkcollege (T)
(4 hours lectures, 2 hours tutorials/computer lab)
Toetsvorm Opdracht (AST), Schriftelijk tentamen (WE)
(Assessment takes place through homework assignments and written exam according to Final = 0.1 x HW1 + 0.1 x HW2 + 0.1 x HW3 + 0.7 x WE only if WE >=5 otherwise Final = WE, where HWi is homework grade for ith homework set, WE final written exam grade.)
Vaksoort bachelor
Coördinator dr. W.P. Krijnen
Docent(en) dr. W.P. Krijnen
Verplichte literatuur
Titel Auteur ISBN Prijs
An Introduction to Generalized Linear Models. Third Edition
(Chapman & Hall/CRC Texts in Statistical Science; 2008)
Annette J. Dobson and Adrian Barnett
Entreevoorwaarden This course builds on Statistics in the second year
Opmerkingen As background literature one can use:
Modern Applied Statistics with S. Fourth Edition. Springer 2002 W. N. Venables and B. D. Ripley
Opgenomen in
Opleiding Jaar Periode Type
BSc Technische Wiskunde 3 semester II a keuzegroep
BSc Wiskunde: General Mathematics  (A. Major track General Mathematics) 3 semester II a keuzegroep
BSc Wiskunde: Statistics and Econometrics 3 semester II a verplicht