Prototype design and development of an automated system for sentiment analysis and meaning analysis of text responses from teaching evaluation surveys at VŠE

This bachelor thesis focuses on the design and development of a prototype automated system for analyzing the sentiment and meaning of textual responses from questionnaires for teaching evaluation at the University of Economics in Prague. The aim of the thesis is to create a functional prototype that enables efficient and accurate processing of student responses using machine learning and large language models (LLM). The thesis first analyses the current state of knowledge in the field of sentiment analysis and text interpretation with a focus on state-of-the-art methods. In the next phase of the research, the system requirements were defined and suitable models were also selected and compared. Subsequently, a prototype was designed and implemented and tested on real data. The prototype provides a user-friendly interface for visualization of results and further data processing options. Finally, recommendations for further improvement of the system and its possible practical use are offered.