Rule−based Categorial Analysis of Unprompted Speech − a Cross−language Study

In this study, we investigated the influence of language specifics in a cross-language task on the automatic segmentation with a self-learning algorithm for the integration of pronunciation rules. The goal of this paper is to present the linguistic and statistic results of a new method to automatically generate pronunciation rules for automatic segmentation of speech the German MAUSER system. MAUSER is an algorithm which generates pronunciation rules independently of any domain dependent training data either by clustering and statistically weighting self-learned rules according to a small set of phonological rules clustered by categories or by re-weighting “seen” phonological rules. For the generation of pronunciation rules the used algorithm does not require any domain dependent training data. By this method we are able to automatically segment cost-effectively large corpora of mainly unprompted speech.