Natural Language Processing and Computational Linguistics Research in Tamil
English to Tamil Machine Translation System (A rule based approach)
Tamil Morphological Tagger, Tamil Robot, Information Retrieval System and Text to Speech Application, Tamil SpellChecker
by Vasu Renganathan, University of Pennsylvania
|Tamil Text to Speech Application | Chat with Tamil Robot ஆயிதழ் அவினி | Tamil Spell Checker with Crowdsourcing option| Tamil Learning Site|
|Read Sangam poems with gloss| Search Sangam Poems| Sangam Quick Search|
Enter simple sentences in English Below. Separate each sentence with a period.
This is an attempt to build a machine translation system purely based on conversion rules between English and Tamil syntactic structures. Neither any statistical method nor pragmatic rules are employed to deal with a vast majority of English sentence types. This system may be construed of as a human-aided domain specific machine translation system, rather than a fully dependent machine translation system. English texts with minimum number of complexities can very well be translated using this system with a possible post-editing process, provided the dictionary contains all of the words from the text.
The English-Tamil syntactic rule base is constructed in such way that additional rules can be added further to improve the efficiency of this system. The principles of Generative grammar, Tamil POS tagger (developed based on Lexical phonology approach of level ordered morphology) and the heuristic power of Visual Prolog are employed heavily in this system to harness the power of digital processing of natural languages.