Context analysis using bigrams

This paper focuses on using bigrams in a topic determination for speech synthesizer. It contains an explanation of a modular architecture for the speech synthesizer and importance of context analysis for customizing and quality enhancement of synthesized speech. The bigram carries information about context and in this work it is shown how to use them to improve the identification of the theme. At the end of this paper results of the topic determination are presented and results are discussed. The whole solution is customized for Slovak language.