Comparison of Methods for the Analysis of Categorical Data with Easy Interpretation

This diploma thesis focuses on introducing suitable methods for analyzing categorical data with easy interpretation. The primary task is to identify, present, and subsequently test these methods on a suitable dataset with categorical data. The work begins with an introduction to the issue of categorical data and an explanation of the concept of explainable AI (XAI). Individual methods for analyzing categorical data with easy interpretation, which can also be easily integrated into the XAI concept due to their decision-making process being easily interpretable, are then presented. Their theoretical basis is adequately summarized, aiming to provide the reader with a comprehensible and structured overview, complete with descriptions of individual methods and approaches. The second part of this thesis is practical, testing individual methods on a dataset with categorical data. The practical part is processed in the Python programming language with the help of statistical and machine learning libraries, and it can also serve as a manual for the reader for applying methods for analyzing categorical data and interpreting their results. In the final part of the thesis, key findings and recommendations for the practical use of methods are summarized. The work provides readers with guidance for selecting the most suitable method of categorical data analysis, depending on specific tasks and requirements, and contributes to a better understanding of the possibilities and limitations of various methods in the field of XAI and their practical significance for the analysis of categorical data.