Asociační pravidla v dotazníkových datech

Abstract

The aim of this bachelor thesis is to design and implement an algorithm for obtaining association rules from questionnaire data. The application is programmed in the C# programming language. It is a desktop application. Development environment used to develop the application is the Visual Studio environment. Preprocessing of questionnaire data plays an important role. The resulting algorithm allows the user not only to select the required attributes from the data, but also to filter the data by groups. There are several interesting characteristics describing the association rules in the output. Finally, the functionality of the algorithm is investigated, experiments are performed that analyze various factors affecting the resulting set of association rules. The theoretical part of this work deals with the knowledge acquisition, its phases. It also manages with association rules and their generation. Subsequently, negative association rules and what characteristics are important in their search. In addition, it deals with the connection of association rules with the GUHA method, which provides a more general view of the types of rules.