Kannada Telugu Machine Aided Translation

Abstract

Machine Translation is one among the first applications of the computer. Since, 1950 there have been several efforts have been made to develop the tools for the language analysis. The languages which have more electronic lexical data on World Wide Web can make use of these lexical tools for building any translation system. But for the languages with limited electronic lexical resources like Kannada and Telugu building a translation system is a challenging and demanding task. Translation is a process of communicating the meaning of text in one language to another. The central goal of this research work is to build a machine translation system from Kannada to Telugu. In the proposed work the source language is Kannada because and target language is Telugu. Human translators understand the meanings and then translate, but computers cannot understand meanings. The translation is done by transferring the structure, which computers are good at, in the hope that meanings preserved. Using appropriate word and sentence structures, it will be easy to produce the meaning full translation. newlineMachine Translation systems can be developed using many approaches. The appropriate approach is selected based on the complexity and structure of the languages. The proposed work uses mainly two approaches for developing a Machine Translation system for Kannada and Telugu language pair and they are Rule based and Neural Machine Translation. Rule based Machine Translation is the appropriate approach for developing a translation system with limited linguistic resources. The rule based approach can be further divided in to three sub approaches they are Dictionary based, Interlingua and transfer based approach. The proposed work uses Dictionary Based and Transfer based approach for developing the translation system newline