Department of Mathematical Sciences, Unit Catalogue 2008/09
MA30084 Generalised linear models
| Credits: 6 |
| Semester: 1|
|Assessment: EX 100%|
|Before taking this unit you must take MA10002 and take MA10003 and take MA10005 and take MA20008 and take MA20035|
Aims: To present the theory and application of generalised linear models, including estimation, hypothesis testing and confidence intervals. To describe methods of model choice and model checking.
On completing the course, students should be able to
* choose an appropriate generalised linear model for a given set of data;
* fit this model using R, select terms for inclusion in the model and assess the adequacy of a selected model;
* make inferences on the basis of a fitted model and recognise the assumptions underlying these inferences and possible limitations to their accuracy.
Generalised linear models: Exponential families, standard form, linear predictors and link functions, deviance. Statement of asymptotic theory for the generalised linear model, Fisher information. Vector and matrix representation.
Model building: Subset selection and stepwise regression methods. Effects of collinearity in regression variables. Model checking including residuals AIC and BIC.