- Dr Karl Schmidt (Reader - Cardiff University)
|Delivery Type||Delivery length / details|
|Lecture||22 x 1 Hour Lectures|
|Assessment Type||Assessment length / details||Proportion|
|Semester Exam||2 Hours Written Exam||100%|
|Supplementary Exam||2 Hours Written Exam||100%|
On successful completion of this module students should be able to:
1. describe the properties of the multivariate Normal distribution;
2. find and identify the distributions of linear and quadratic forms in Normal variates;
3. formulate a given situation as a (matrix) linear model;
4. analyse data from experiments modelled in this way;
5. construct confidence intervals / regions for linear combinations of parameters and for unknown population variance.
6. carry out simple linear hypothesis tests based on ANOVA.
7. calculate leverages and residuals, and use them for outlier detection.
The Linear Statistical Model encompasses a variety of elementary statistical techniques such as linear regression, design models, ANOVA, etc, and much more besides. This module introduces the general matrix formulation of the linear model, and demonstrates the neatness of its systematic application to a wide range of statistical problems.
To introduce the scope and breadth of linear matrix modelling.
2. GENERAL LINEAR MODEL OF FULL RANK: Formulation. Ordinary Least Squares estimator and normal equations. The assumption of independent homoscedastic errors. The BLUE and Gauss-Markov Theorem. Consideration of the design matrix. The hat matrix and leverage.
3. FURTHER INFERENCE IN THE FULL RANK CASE: Estimation of population variance. Confidence regions and intervals. Consideration of outliers: residuals and leverage. Tests of linear hypotheses.
This module is at CQFW Level 6