Enhanced Rule-Based Phonetic Transcription for the Romanian Language
Mihai Alexandru Ordean, Andrei Şaupe, Mihaela Ordean, Melania Duma, Gheorghe Cosmin Silaghi · 2009
This paper presents a rule-based approach for the phonetic transcription of the Romanian language. We integrate this phonetic analysis in the text processing component of a text-to-speech system for Romanian. Grapheme-to-phoneme rules are constructed based on expert information from DOOMII dictionary. In the cases when rules are useless, we employed decision trees constructed on engineered training sets to help the classifiers to learn the language exceptions.