MorphoClass: а System for Recognition

Abstract: A system for recognition and morphological classification of unknown words for German is described and evaluated. It takes raw text as input and outputs a list of the unknown nouns together with a hypothesis about their possible morphological class and stem. MorphoClass exploits global information (ending-guessing rules, maximum likelihood estimations, word frequency statistics), morphological properties (compounding, inflection, affixes) and external knowledge (lexicons, German grammar information etc.). Key words: knowledge acquisition, computational linguistics, morphological analysis, natural language processing, part-of-speech guessing, compounds analysis. 1. SYSTEM OVERVIEW The MorphoClass system accepts raw text as input and produces a list of unknown words together with hypotheses about their stem and morphological class. We define the stem as the common part shared by all inflected forms of the base while the morphological class describes both the word gender and the inflexion rules the word follows when changes by case and number. Our morphological classes follow the one used under the DBR-MAT project — a German-Bulgarian-Romanian Machine Translation (v. Hahn & Angelova, 1996), and given in Bulgarisch-Deutsch Worterbuch (Dietmar & Walter, 1987). MorphoClass solves the problem as a sequence of subtasks including: unknown words identification, nou