Generic Morphological Analysis Shell

Abstract

Morphological analyzers are mainly available for major languages like English, Spanish and French. This paper describes a generic shell which can be used to develop morphological analyzers for different languages particularly minority languages. The shell uses finite state transducers with feature structures to give the analysis of a given word. The most significant aspect of the shell is the integration of paradigms with augmented FSTs. The simple format in which a linguist may provide data in the form of dictionaries, paradigm and transition tables is the other important feature of the shell. The shell has been experimented on Hindi, Telugu, Tamil and Russian languages.