Analysis of sanskrit text: Parsing and semantic relations

Abstract

In this paper, we are presenting our work towards building a dependency parser for Sanskrit language that uses deterministic finite automata(DFA)for morphological analysis and 'utsarga apavaada' approach for relation analysis. A computational grammar based on the framework of Panini is being developed. A linguistic generalization for Verbal and Nominal database has been made and declensions are given the form of DFA. Verbal database for all the class of verbs have been completed for this part. Given a Sanskrit text, the parser identifies the root words and gives the dependency relations based on semantic constraints. The proposed Sanskrit parser is able to create semantic nets for many classes of Sanskrit paragraphs( anuchhed). The parser is taking care of both external and internal sandhi in the Sanskrit words.