Machine learning of European Portuguese grapheme-to-phone conversion using a richer feature set

Abstract

In this study evaluation of two self-learning methods (MBL and TBL) on European Portuguese grapheme-tophone conversion is presented. Combinations (parallel andcascade) of the two systems were also tested. The usefulness of using syllable related information in machine learning approaches is also investigated. Systems with good performance were obtained both using a single self-learning method and combinations. Best performance was obtained with MBL and the parallel combination. The use of syllable information contributes to a better performance in all systems tested, being the effect significant statistically. Our best machine based systems present Word Error Rate and Mean Normalized Levenshtein Distance similar to those recently obtained for German when using similar features.