State of the Art in Computational Morphology

From the point of view of computational linguistics, morphological resources are the basis for all higher-level applications. This is especially true for languages with a rich morphology, such as German or Finnish. A morphology component should thus be capable of analyzing single word forms as well as whole corpora. For many practical applications, not only morphological analysis, but also generation is required, i.e., the production of surfaces corresponding to speci?c categories. Apart from uses in computational linguistics, there are also numerous practical - plications that either require morphological analysis and generation or that can greatly bene?t from it, for example, in text processing, user interfaces, or information - trieval. These applications have speci?c requirements for morphological components, including requirements from software engineering, such as programming interfaces or robustness. In 1994, the First Morpholympics took place at the University of Erlangen- Nuremberg, a competition between several systems for the analysis and generation of German word forms. Eight systems participated in the First Morpholympics; the conference proceedings [1] thus give a very good overview of the state of the art in computational morphologyfor German as of 1994.

Contenu

Using Ranked Semirings for Representing Morphology Automata.- JSLIM - Computational Morphology in the Framework of the SLIM Theory of Language.- HFST Tools for Morphology - An Efficient Open-Source Package for Construction of Morphological Analyzers.- fsm2 - A Scripting Language for Creating Weighted Finite-State Morphologies.- Morphisto: Service-Oriented Open Source Morphology for German.- Morphosyntactic and Semantic Analysis of Text: The MPRO Tagging Procedure.- Word Manager.- Morphological Analysis Using Linguistically Motivated Decomposition of Unknown Words.- Corpus-Based Lexeme Ranking for Morphological Guessers.