In many statistical problems, we have to work with a considerable number of variables. Factorial methods allow us to reduce this number of variables in order to get a graphical representation of the data. In this thesis, we will study two methods: the principal component analysis and the factorial discriminant analysis. We will develop them in the classical case and in the symbolic case and we will apply them to the analysis of symbolic data with the Sodas software.
|Date of Award||27 Jun 2006|
|Supervisor||Andre Hardy (Supervisor), MARCEL REMON (Jury) & Jean Paul Rasson (Jury)|