Generalised Linear Models

The theory for linear normal models is looked at and applied to regression and analysis of variance. Furthermore the topics of binary variables logistic regression, log-linear models, contingency tables and life time analysis are treated.

After completed course, the students are expected to be able to:

• Identify probability distributions belonging to an exponential family and adapt a description as a generalized linear model.
• Present the general theory of exponential families of distributions.
• Describe numerical procedures for estimation in generalized linear models.
• Recognize linear normal models and apply general test procedures to these models.
• Explain the proofs of important theorems in probability theory utilized in test procedures in linear normal models and in generalized linear models.
• Analyze data sets following Poisson or binomial distributions.
• Estimate parameters and test hypotheses in generalized linear models by means of statistical software.

Irregular