Vybrané vícerozměrné metody pro analýzu kategoriálních dat

The diploma thesis is focused on the introduction of selected multidimensional methods for the analysis of categorical data. Emphasis is placed primarily on hierarchical clustering algorithms and correspondence analysis, which is a suitable visualization tool for associations between variables and their categories. The methods are accompanied by exploratory data analysis. To demonstrate the methods, a real data set is used in the work, on which selected procedures and methods are illustrated. In the end important findings about the features of the data set, which were reached based on the analysis, are summarized. These methods are processed using R software, for two main reasons. The R software includes a freely available package of newly designed similarity measures for hierarchical clustering of nominal data, and these options are not yet implemented in commercial software. The second main reason is the introduction of the possibilities of application of correspondence analysis in freely available software widely used not only in the academic sphere, but also in practice. The aim of the diploma thesis is to introduce these methods and their application in R software. Using these methods, the work seeks to explain the relationship between the elected political party and other aspects.