Bidirectional Analyzer and Generator Tool for Kannada Nouns

Abstract

This paper explores a computational model for Kannada nouns. This is a single tool performing analysis and generation in reverse direction. Kannada has rich morphology with inflections -vibhakti, derivations and compound formation. In depth study of words and its structures is done using this tool. Through study of noun from computational linguistics point of view is important. An exhaustive version of the morphological analyzer and generator tool for nouns using FST is described here. Morphological analyzer/generator is designed to interface with a knowledge-based machine translation (MT) system. Our system uses a hierarchy (tree) structure to relate various morphological features. Morphological analyzer/generator for morphologically complex and agglutinative language like Kannada is highly challenging. The major types of morphological process like inflection, derivation, and compounding are handled in this system. All these developments have worked on the basis of syntactic structure of Kannada language. A robust analyzer is useful for language learning as well as for machine translation applications. The tool is tested against 10000 words and has 98% recognition success rate.