Analýza topiků v textech

Abstract

In this diploma thesis, we discuss text processing and analysis by using topic modeling. The main goal of theoretical part is to find and describe current methods of topic modeling. Different topic evaluation methods and text preprocessing approaches are also discussed. Practical chapters are dedicated to application of chosen topic models on texts. In experiments, we compare various methods of preprocessing and how they affect topic models, we analyze topics of best models and we visualize one of them as a network.