Word Reordering using Rule based Parser in Telugu to English Machine Translation

Abstract

In Machine translation Word Reordering is an important intermediary stage, where the source language sentences with disordered words are oriented into the grammatical structure of target language. The main focus is imposed on the reordering of given words into the basic structure of target language using parser. A parser is employed on the POS tagged words, which uses the rule bank to orient the words into correct form. The rule bank is a collection of rules that consists of the syntax of target language. The parser with the set of rules, reorders the source language structured sentence into target language structure. The paper mainly focuses on the implementation of reordering for Telugu to English Translation where the SOV (Subject-Object-Verb) order of source language Telugu is converted into SVO (Subject-Verb-Object) order of target language English. The output of Reorder stage is presented in the format of a parse tree that forms the basis for the further post processing required.